Full text of "CREAMS : a field scale model for Chemicals, Runoff, and Erosion from Agricultural Management Systems"

See other formats

Historic, archived document

Do not assume content reflects current
scientific knowledge, policies, or practices.

\ /

cm x « >< w >< ® >< w x

<0301 Baltimore Bwd
Sto.MD 20705-2351

VSDA, National Agricultural Library

i y\QA Baltimore Blvd

. 2351

CREAMS

A Field Sc^Ie MQdgJls>r—
Chemicals, Runoff, and Erosion From
Agricultural Management Systems

MANAGEMENT INPUT

NATURAL
INPUT

PRECIPITATION
(RAIN, SNOW)

TEMPERATURE

LAND USE

CULTURAL
PRACTICES

PLANT
NUTRIENTS

PESTICIDES

USDA, National Agricultural Library

NAL Bldg

10301 Baitimore Blvd

Beitsviile, MD 20705-2351

■EVAPOTRANSPIRATION

SURFACE RUNOFF

PERCOLATION

EROSION/
SEDIMENTATION

DISSOLVED
CHEMICALS

ADSORBED
CHEMICALS

SSs\ UNITED STATES
Uy)) DEPARTMENT OF
^^ AGRICULTURE

CONSERVATION
RESEARCH REPORT
NUMBER 26

PREPARED BY
SCIENCE AND
EDUCATION
ADMINISTRATION

740738

AGS
CREAMS

A Field Scale Model for

Chemicals, Runoff, and Erosion From

Agricultural Management Systems

VOLUME I. MODEL DOCUMENTATION

U.S. Department of Agriculture
Nations ' ^ricultural Library
Lenc: -eh

Baltsville, Maryland 20705

CONTENTS
Chapter Page

1 Introduction- -------------------------- l

— W. G. Knisel and A. D. Nicks

2 Simulation of the surface water hydrology ------------ 13

--R. E. Smith and J. R. Williams

3 A model to estimate sediment from field-sized areas ------- 36

--G. R. Foster, L. J. Lane, J. D. Nowlin,
J. M. Laflen, and R. A. Young

4 The nutrient submodel ---------------------- 65

--M. H. Frere, J. D. Ross, and L. J. Lane

5 The pesticide submodel- --------------------- 88

--R. A. Leonard and R. D. Wauchope

6 Sensitivity analysis- ---------------------- l 13

--L. J. Lane and V. A. Ferreira

ABSTRACT

Knisel, Walter G., editor. CREAMS: A Field-Scale Model for Chemicals, Runoff ,
and Erosion from Agricultural Management Systems. U.S. Department of Agricul-
ture, Conservation Research Report No. 26, 640 pp., illus.

This publication describes a mathematical model developed to evaluate non-
point source pollution from field-sized areas. CREAMS consists of three compo-
nents: hydrology, erosion/sedimentation, and chemistry. The publication is
presented in three volumes: Volume I, model documentation, describes the con-
cepts of each of the model components; Volume II, user manual, describes the
model application and selection of parameter values; Volume III, supporting
documentation, provides additional data and parameter information.

Keywords: hydrology, erosion, sediment transport, plant nutrient transport,
pesticide transport, mathematical model, nonpoint source pollution, agricul-
tural management.

PREFACE

Section 208 of PL 92-500, the 1972 amendments of the Clean Water Act,
placed emphasis on nonpoint-source pollution. Planning required by this legis-
lation needed methods to assess nonpoint-source pollution under various manage-
ment practices for selecting Best Management Practices (BMP's) to reduce non-
point pollution to acceptable levels.

Expertise by the staff of the Science and Education Administration-Agri-
cultural Research (SEA-AR) in soil and water management research, along with
the high priority needs of action agencies, prompted SEA's National Program
Staff (NPS) to develop plans for a concerted national effort to assemble mathe-
matical models for evaluating nonpoint-source pollution. Staff scientists met
with planners in action agencies to determine their needs for such models. A.
R. Robinson, D. A. Farrell, and J. Lunin, all of the NPS, and J. C. Lance, tem-
porarily assigned to this staff, planned the mechanism for a national project.
The plans were approved by C. W. Carlson, Associate Administrator of Agricul-
tural Research, and a request was made to T. W. Edminster, Administrator of
Agricultural Research, for unassigned program funds to initiate the project.
T. W. Edminster and R. J. McCracken, then the Assistant Administrator, made
funds available for this important project.

The project coordinator and Technical Work Group met with the Steering
Committee (NPS scientists) at Beltsville, Md., in October 1977, to initiate a
national project to develop mathematical models for evaluating nonpoint-source
pollution. On February 14-16, 1978, a workshop was held at Arlington, Tex., to
assemble SEA-AR scientists interested in participating in this project. The
workshop was to:

...(1) review, refine, and adopt an approach, (2) select group
leaders (lead scientists), (3) plot the course of action and
set the time table..., (4) assign tasks (to) investigate spe-
cific components of the system to be considered...

To develop a model quickly, participants at the workshop determined that
existing physically based models, or those that could be readily modified and
improved, would be assembled into a package to estimate runoff, sediment, plant
nutrient, and pesticide movement in a field.

Lead scientists identified for the four components are:

Hydrology A. D. Nicks, Chickasha, Okla.

Erosion G. R. Foster, Lafayette, Ind.

Plant nutrients — M. H. Frere, Chickasha, Okla. (New Orleans, La.)

Pesticides R. A. Leonard, Athens, Ga.

The hydrology component was further represented by two options lead by R.
E. Smith, Fort Collins, Colo., and J. R. Williams, Temple, Tex. J. D. Nowlin,

CONTENTS

Page

Volume I. Model documentation- ------------ ------- i

Volume II. User manual ------------------------ 159

Volume III. Supporting documentation- ----------------- 377

Trade names are used in this publication solely for the purpose of provid-
ing information. Mention of a trade name does not constitute a guarantee
or warranty of the product by the U.S. Department of Agriculture or an en-
dorsement by the Department over other products not mentioned.

This publication reports research involving pesticides. It does not con-
tain recommendations for their use, nor does it imply that the uses dis-
cussed here have been or remain registered. All uses of pesticides must
be registered by appropriate State and/or Federal agencies before they can
be recommended.

Issued May 1980

Department publications contain public information. They are not copyrighted
and may be reproduced in whole or in part with or without credit.

Purdue University, Lafayette, Ind., working with SEA-AR under cooperative
agreement, programmed the model concepts.

The lead scientists drew upon material provided by other contributors to
develop and document the model. These contributors are acknowledged throughout
the publication.

Scientists in SEA-AR worked together to assemble state-of-the-art mathe-
matical models to evaluate nonpoint-source pollution for field-scale areas,
pesults of these efforts have culminated in an operational continuous simula-
tion model. This publication documents and provides a user manual for the mod-
el named CREAMS - a field scale model for Chemicals, Runoff, and Erosion from
Agricultural Management Systems.

The CREAMS model was developed using units common for the individual com-
ponents. That is, customary units in the hydrology and erosion/sedimentation
fields are English units, whereas metric units are common in chemistry. Rain-
fall data available from the National Weather Service and SEA-AR are reported
in inches. Temperature data are generally in degrees Fahrenheit. Runoff, per-
colation, soil water, and evapotranspiration are generally reported in inches.
Erosion/sedimentation data are generally reported in pounds per acre or tons
per acre. Plant nutrient and pesticide losses are reported in milligrams per
liter and kilograms per hectare. Model input, output, and operations were
structured accordingly. Although the CREAMS model has the potential for inter-
national use, the principal users will be action agencies and consulting firms
in the United States and the model, therefore, contains mixed or customary
units. Users are cautioned, however, against indiscriminately modifying model
components without a complete understanding of the units of operations. This
version will be improved over the next several months to provide a more compre-
hensive model and will incorporate consistent English or metric units for user
option specification.

The purpose of this publication is to provide a complete package for po-
tential users of the model. It is divided into three main divisions. Volume
I, model documentation, presents the concepts of model components. Volume II,
user manual, provides information on selection of parameter values and model
operation. Volume III, supporting documentation, provides the user additional
information to obtain parameter values.

Results of sensitivity analysis for each component are included in the
publication to indicate effects of errors in parameter estimation. This en-
ables the user to be aware of potential difficulties resulting from inaccura-
cies in individual parameters.

The results of considerable testing of the components using data available
from SEA-AR research locations are included. Users should be aware of two sig-
nificant points: (1) Statements of model accuracy in the publication are made
realistically based upon the scientist's evaluation of the mathematical repre-
sentation of the real-world system and his scientific knowledge of the range
and confidence in parameter estimation, and (2) ranges of conditions considered
appropriate for application of the model are given in the publication.

Magnetic tapes of the computer model can be furnished to anyone interested
in using the model. A user, in turn, can send a magnetic tape to the project

coordinator and the program will be taped along with a set of test data, param-
eter values, and summary output. These data will enable users to be sure the
model is operating properly on their respective computer systems. A tape can
be generated for CDC or IBM computers and the user should specify the system
when requesting the program.

Walter G. Knisel , Jr.

Project Coordinator

USDA-SEA-AR

442 East Seventh Street

Tucson, Ariz. 85705

ACKNOWLEDGMENTS

The authors acknowledge T. W. Edminster, R. J. McCracken, and C. W.
Carlson for their confidence in and support of the scientists working on the
modeling project.

National Program Staff scientists who make up the Steering Committee for
the project, D. A. Farrell, J. Lunin, J. C. Lance, and A. R. Robinson, are
recognized for their technical support.

The Technical Work Group, D. G. DeCoursey, E. T. Engman, L. D. Meyer, M.
H. Frere, and R. A. Leonard, is recognized for planning and conducting a work-
shop at Arlington, Tex., to initiate the modeling effort.

K. G. Renard is acknowledged for providing support staff assistance to the
project. Virginia Ferreira, Karen Mellor, Sue Schield, John Rocha, and Bob
Wilson are recognized for their respective assistance in computer terminal
operations, typing, and drafting.

W. E. Modenhauer, SEA-AR, and G. W. Isaacs, Agricultural Engineering De-
partment, Purdue University, Lafayette, Ind., are recognized for providing J.
D. Nowlin for computer programming assistance.

The following scientists participated in the workshop at Arlington, Tex.,
in 1978 and contributed ideas, direction, and data as well as completed task
assignments that led to the model development.

G .

Alonso, Oxford, Miss.
Barnett, Watkinsville, Ga
Bonta, Coshocton, Ohio
Bowie, Oxford, Miss.
Chery, Athens, Ga.
Cooley, Phoenix, Ariz.
DeCoursey, Oxford, Miss.
Engman, Beltsville, Md.
Foster, Lafayette, Ind.
Frere, Chickasha, Okla.
Hamon, Coshocton, Ohio
Hanson, Boise, Idaho
Hjelmfelt, Columbia, Mo.
Kissel , Temple, Tex.
Knisel , Tucson, Ariz.
Laflen, Ames, Iowa
Lance, Beltsville, Md.
Lane, Tucson, Ariz.
Larson, St. Paul , Minn.
Leonard, Athens, Ga.

Murphree, Oxford, Miss.
Mutchler, Oxford, Miss.
Nash, Beltsville, Md.
Neff, Sidney, Mont.
Nicks, Chickasha, Okla.
Onstad, Morris, Minn.
Owens, Coshocton, Ohio
Piest, Columbia, Mo.
Pionke, Univ. Park, Penn.
Rawls, Beltsville, Md.
Richardson, Temple, Tex.
Ritchie, Beltsville, Md.
M. Romkens, Oxford, Miss.
Saxton, Pullman, Wash.
Seely, Chickasha, Okla.
Smika, Akron, Colo.
Smith, Fort Collins, Colo,
Smith, Durant, Okla.
Spencer, Riverside, Calif,
Timmons, Morris, Minn.

D. K. McCool , Pullman, Wash. R. D. Wauchope, Stoneville, Miss.

L. L. McDowell, Oxford, Miss. J. R. Williams, Temple, Tex.

R. G. Menzel , Durant, Okla. G. H. Willis, Baton Rouge, La.

L. D. Meyer, Oxford, Miss. R. A. Young, Morris, Minn.

J. B. Burford and Jane DeLashmutt of the Water Data Laboratory, Belts-
ville, Md., helped obtain and format data for testing the hydrologic components
of the model .

USDA-Soil Conservation Service personnel critiqued the model in a technol-
ogy transfer workshop at the South Technical Service Center, Fort Worth, Tex.
John Burt, Gary Margheim, E. C. Nicholas, and S. J. Robbins helped arrange the
workshop and provided input to improve the model. Margheim obtained SCS funds
for the SEA-AR model testing and technology transfer.

B. C. Dysart, III, and R. C. Warner, Environmental Systems Engineering De-
partment, Clemson University, Clemson, S.C., are acknowledged for their com-
ments and suggestions on the erosion component. Their application in South
Carolina enabled improvements of the erosion component.

CREAMS: A FIELD SCALE MODEL FOR CHEMICALS, RUNOFF, AND EROSION FROM
AGRICULTURAL MANAGEMENT SYSTEMS

VOLUME I. MODEL DOCUMENTATION

Chapter 1. INTRODUCTION
W. G. Knisel and A. D. Nicks-/

Under the Federal Water Pollution Control Act Amendments of 1972, Public
Law No. 92-500, the Administrator of the Environmental Protection Agency (EPA),
in cooperation with other agencies, provides guidelines for identifying and
evaluating nonpoint sources of pollutants. The U.S. Department of Agriculture
(USDA) is one of the cooperating agencies. The USDA-Soil Conservation Service
(USDA-SCS) has the technical responsibility for evaluating nonpoint source pol-
lution and implementing Best Management Practices (BMP's) to limit nonpoint
source pollution to an acceptable level. The Science and Education Administra-
tion-Agricultural Research (SEA-AR), as the research agency of USDA, has obli-
gations for research to meet the needs of SCS and EPA.

Scientists in SEA-AR (formerly Agricultural Research Service) by request
of EPA prepared a two-volume document on control of potential water pollutants
from cropland. The two volumes, published in 1976, include information on the
basic principles of control of specific pollutants (28, 29). A list of BMP's
was included in volume I (28). Although simple models for estimating annual
values of runoff, percolation, erosion, plant nutrient, and pesticide losses
were given in volume II (29), the BMP's were not quantified. Management prac-
tices are site-specific. Stewart and others, (28) stated: "Because of the
variation of climate, soils, and agricultural practices throughout the United
States, no single group of control measures can be used for every region, nor
will the regional information printed herein be accurate for all areas within
the region."

SEA-AR recognized the need for development of physically based mathemati-
cal models to make the next logical step beyond the Stewart and others reports
(28, 29) , and in 1978 scientists began a concerted effort to assemble such mod-
els. Since management practices are applied on a farm or field basis, it was
thought that the size range to be considered should be the field scale. Figure
1-1 shows a schematic representation of a field with natural and management in-
put and the associated water, sediment, and chemical output.

1/ Hydraulic engineer, USDA-SEA-AR, Tucson, Ariz., and hydraulic engineer,
USDA- SEA-AR, Chickasha, Okla., respectively.

MANAGEMENT INPUT

Figure 1-1. — Flow chart of system for evaluating nonpoint
source pollution.

A question arose immediately: What size is a field? The physical size of
farm fields varies from a few acres in ridge and valley provinces to a few tens
of acres in the Corn Belt to a few hundreds of acres in the Wheat Belt and
western rangelands. Such a size range required some arbitrarily imposed con-
straints. Thus, a field herein is defined as a management unit having (1) a
single land use, (2) relatively homogeneous soils, (3) spatially uniform rain-
fall, and (4) single management practices, such as conservation tillage or ter-
races. This definition allows different physical sizes in different climatic
regions and Land Resource Areas (LRA's).

To achieve the goal of model assembly in a year, state-of-the-art models
were assembled and/or modified. Criteria for the model were: (1) the model
must be physically based and not require calibration for each specific applica-
tion, (2) the model must be simple, easily understood with as few parameters as
possible and still represent the physical system relatively accurately, (3) the
model must estimate runoff, percolation, erosion, and dissolved and adsorbed
plant nutrients and pesticides, and (4) the model must distinguish between man-
agement practices.

Although hydrology is only one component of the total system, water is
the principle element; it causes erosion, carries chemicals, and is an uncon-
trolled natural input. Each climatic region and physiographic area has its own
characteristics that affect the response of the system. These varied condi-
tions must be kept in mind when considering wide-scale applicabilty of a model.
Figure 1-2 is a generalized schematic representation of the water balance for
different areas of the United States. The width of bars in the figure are
drawn to scale to show the relative magnitude of each component among the five
regions. In the Southeast Coastal Plain, rainfall averages about 50 inches
(1,270mm), and evapotranspiration is about 35 inches (890 mm). Approximately
80 percent of the water that ultimately reaches streamflow has at one time been
subsurface flow. That is, about 12 inches (305 mm) of total streamflow comes
from subsurface flow (24). Only 3 inches (76 mm) comes from direct overland

SOUTH CENTRAL US

NBELT

NORTHWEST RANGELAND

SOUTHWEST RANGELANO

Figure 1-2. — Schematic representation of water
balance for selected locations in the United
States.

flow. Deep percolation to regional groundwater is negligible. In the south
central United States, average rainfall is about 34 inches (864 mm), and runoff
is about 8 inches (203 mm). For all practical purposes, there is no subsurface
flow or deep percolation, and evapotranspiration is about 26 inches (661 mm).
In the semiarid Southwest, precipitation is about 13 inches (330 mm), with only
0.5 inch (13 mm) of surface runoff and negligible groundwater recharge or sub-
surface flow. Snow is a part of the precipitation input for the Corn Belt and
Northwest rangelands. Subsurface flow and groundwater recharge are significant
components in the Corn Belt. Dissolved chemicals may be an important potential
nonpoint source pollutant in the Coastal Plain and in the Corn Belt. Sediment
and adsorbed chemicals may be the major pollutants in the Western rangelands
and the South Central areas, as well as the Corn Belt. Although these repre-
sentations are generalized, they indicate the varied conditions that a nonpoint
source pollution model must be capable of considering.

The system represented in figure 1-1, the conditions represented in figure
1-2, the model criteria, and the constraints of field size were guidelines used
in the development and testing of CREAMS. This publication documents the model
and provides a user manual to aid in selection of parameter values to run the
model. CREAMS is the first step beyond the Stewart and others reports (28,
29) , and is preliminary to a basin scale model.

The general logic of the model is that hydrologic processes provide the
transport medium for sediment and agricultural chemicals. Therefore, the hy-
drologic component provides input to the other model components. The erosion/
sediment yield component in turn provides estimates of sediment yield and silt/
clay/organic matter enrichment to be used in the chemical transport components.
The documentation generally follows this logic, with evaluation included in
each section. A separate section gives results of sensitivity analysis of the
model parameters.

DEVELOPMENT OF NONPOINT SOURCE POLLUTION MODELS

Hydrologists have long used models to depict relationships between such
hydrologic variables as rainfall, runoff, evapotranspiration, and infiltration.
These were generally graphical representations or regression equations that
could be solved easily with desk calculators. The relationships like the ra-
tional formula Q = CIA (13) often were gross simplifications of complex proces-
ses.

In the 1950' s, the USDA-SCS recognized the need for a more comprehensive
model to estimate runoff from rainfall as a function of soil, vegetation, and
antecedent moisture, and developed the SCS curve number (SCSCN) model. The ba-
sic model is still being used at present by SCS (30) . This model related storm
runoff to storm rainfall, and was used to estimate runoff in the report of
Stewart and others (_29) .

Wischmeier and Smith (33) analyzed many years of plot data to develop the
Universal Soil Loss Equation (USLE) for estimating gross erosion by water. The
USLE, a relatively simple regression equation, is presently being used by many
agencies and consultants, and it was updated recently (34).

Development of electronic data processing equipment eliminated the time-
consuming repetitive hand calculations necessary for analyses of large volumes
of data. Also, scientists may now formulate more complex conceptual models and
solve more complex equations with the computer than was possible earlier. Mod-
el proliferation began in the late 1950's, and continues to the present. These
models cover a range of sophistication and mathematical complexity. Models
range from deterministic to stochastic, with various combinations in between.
The models were all developed for specific purposes that ranqe from analyses of
data to extrapolation of data to some future condition. These specific pur-
poses include prediction of runoff from rainfall, estimation of erosion from
rainfall, projection of downstream sediment yield from field erosion processes
within a watershed, and so forth.

In 1962, Crawford and Linsley (_4) published one of the earliest computer
hydrologic simulation models. The model became widely known as the Stanford
watershed model. It uses conceptual simplifications for physical processes of
overland flow, interflow, upper zone soil water storage, lower zone soil water
storage, deep percolation, groundwater storage, and evapotranspiration to esti-
mate streamflow from rainfall records (_5_ ) . The model requires calibration to
specific watershed conditions and was primarily intended to show effects of wa-
tershed changes on streamflow.

The Stanford watershed model (SWM) became the basis for numerous studies,
and several scientists have made revisions, particularly in optimization proce-
dures for calibration (_20, 25J . More recently, the Stanford model has been
used as the basic hydrologic component for field-scale, water-quality models
(_3 ) . The basic concepts of the model were retained with internal revisions,
but calibration of the model to specific fields is still required.

Glymph and Holtan (15) developed an infiltration-based hydrologic model,
known as the USDAHL (U.1T Department of Agriculture Hydrograph Laboratory)

model, to estimate streamflow using a concept of soil zones on the watershed
landscape. Snowmelt, separation of flow regimes, and ground water contribu-
tions to streamflow have been incorporated recently (17, 18) .

Passage of the Federal Water Pollution Control Act Amendments, Public Law
92-500, (commonly known as the Clean Waters Act) in 1972 created an awareness
by many agencies and consultants for models to simulate processes affecting wa-
ter quality. More specifically, Section 208 of the Clean Waters Act specified
that by October 1978 the States would have completed plans for limiting stream-
flow pollution from nonpoint sources, particularly agriculture. This specifi-
cation emphasized the need for mathematical models to evaluate nonpoint-source
pollution and consider BMP's to reduce the pollution (_6) . All these models,
that is, SCSCN, USLE, USDAHL, and the SWM, were used later as basic components,
with or without modification, for water-quality models. There was little pre-
cedent for chemical transport models, especially for upland areas, although
diffusion models had been applied in river-channel systems. Since water is the
carrier of sediment and chemicals, most water quality models were developed by
selecting a hydrologic model, and "piggy-backing" sediment and chemistry com-
ponents to produce a model package.

Hydrocomp, under contract with the Environmental Protection Agency (EPA),
developed the Pesticide Transport and Runoff (PTR) model (3). A revision of
the SWM (_5) became the hydrologic component of the PTR model. The sediment
loss component of PTR consists of a part of Negev's equation for sediment de-
tachment and transport (23). Although Negev simulated the entire sheet, rill,
and channel erosion, the PTR model only uses the sheet and rill erosion compon-
ents which include the detachment and transport of soil particles by overland
flow. Pesticide simulation includes the process of adsorption/desorption to
determine the division between the sediment and water phases of runoff. Vola-
tilization of pesticides is considered along with degradation, which is repre-
sented by a first-order, decay-type relation with time. Plant nutrients were
not considered in the PTR model.

Frere, Onstad, and Holtan (12) developed an agricultural chemical trans-
port model (ACTMO) based on the USDAHL model (_18 ) . The erosion/sediment trans-
port component of ACTMO is a modification of the USLE to reflect both rainfall
and runoff erosivity and transport processes (11) . The erosion component esti-
mates the contribution of rill and intern' 11 sources to sediment load. The
chemical component of ACTMO included pesticide and nitrate options. The pesti-
cide option treated adsorption, breakdown, and movement processes, \lery little
field data were available to validate the proposed relationships. The nitrate
option considered mineralization, plant uptake, and movement processes.

Bruce and others (_2 ) , developed a parametric model for water-sediment-
chemical (WASCH) runoff for single storm events. The hydrologic component con-
sists of three functions: a retention function, a characteristic function, and
a variable state function (26). Two-stage convolution is used to produce non-
linear watershed response. The sediment component of WASCH considers the rill/
intern' 11 erosion concepts developed by Foster and Meyer (10), but uses erosion
and routing functions for both rill and interrill erosion. Sediment transport
capacity in the WASCH model is a function of overland flow discharge rather
than velocity. The chemical component of WASCH considers only pesticides and
does not treat plant nutrients. The pesticide model is a single mathematical

expression relating pesticide runoff to rill and interrill sediment with ex-
traction and enrichment factors.

Donigian and Crawford (_7) modified, tested, and further developed the PTR
model, and these revisions resulted in the Agricultural Runoff Management (ARM)
model. Although the model was revised, the original basic components were the
same, that is, SWM for the hydrology component, and Negev's equations for the
sediment component. A plant nutrient component was incorporated into the new
version.

Donigian and Crawford (8) developed a Nonpoint Source Pollutant loading
(NPS) model to simulate polTutant contributions to stream channels from non-
point sources. NPS considers a maximum of five pollutants from each of a maxi-
mum of five separate land use categories. The hydrology and erosion components
are identical with those in ARM (_7 ) . The water quality component relates
pollutants to sediment by specifying pollutant strength or potency factors.
NPS does not have a component for channel processes, but simulates loads of
pollutants reaching the stream channels.

Williams and Hann (3_2) developed a basin scale model to consider surface
runoff, sedimentation, and plant nutrients. The hydrologic component is a
modification of the SCS curve number model. The USLE was modified for the
erosion component by replacing the rainfall energy term with a product of storm
runoff volume and peak rate of discharge raised to a power. Subsurface or
baseflow is not considered by the model. The plant nutrient component of the
model considers both organic and inorganic nitrogen and denitrif ication,
immobilization, and mineralization processes. Nitrogen fertilization, nitrogen
in rainfall, and nitrogen from crop residue are inputs to the basin soils,
while plant uptake and nitrate leaching were simulated to remove nitrogen from
the soils. The phosphorus component of the nutrient model considered only that
portion adsorbed to soil particles. Both the nitrogen and phosphorus
components use enrichment ratios to develop loading functions. The model
routes runoff, sediment, nitrogen, and phosphorus to the basin outlet. Linear
programming techniques are used to select an alternate management practice.

Gianessi, Pleskin, and Young (14) developed a water pollution network mo-
del, referred to as the RFF model [Resources for the Future), to link sources
of pollutants to concentrations in water bodies throughout the Nation. The wa-
ter network identifies 1,051 node points along rivers of the United States to
correspond with U.S. Geological Survey (USGS) gaging station locations. Each
county in the United States is assigned to at least one node. The average
distance between nodes is 66 miles. Streams were classed by ranges of mean
discharges, and USGS periodic stream gaging measurements at the nodes are used
to determine velocity at the nodes. The RFF model emphasizes pollutants,
including sediment, which are input at node points and are assumed input
uniformly between nodes. Loading functions, on a county basis, are obtained
from McElroy and others (21) . Sediment from construction, forestry, and mining
activities is obtained by prorating national estimates to each county based on
the county's share of employment in these activities and weighted by an
estimate of runoff. The RFF model is basically a routing technique for 66-mile
river reaches with generalized loadings of pollutants without identification of
conservation systems on less than a county basis.

The CREAMS is a physically based, daily simulation model that estimates
runoff, erosion/sediment transport, plant nutrient, and pesticide yield from
field-sized areas. The hydrologic component consists of two options. When
only daily rainfall data are available to the user, the SCS curve number model
is used to estimate surface runoff. If hourly or breakpoint rainfall data are
available, an infiltration-based model is used to simulate runoff. Both meth-
ods estimate percolation through the root zone of the soil. The erosion com-
ponent maintains elements of the USLE, but includes sediment transport capacity
for overland flow. A channel erosion/deposition feature of the model permits
consideration of concentrated flow within a field. Impoundments are treated in
the erosion component also. The plant nutrient submodel of CREAMS has a nitro-
gen component that considers mineralization, nitrification, and denitrif ication
processes. Plant uptake is estimated, and nitrate leached by percolation out
of the root zone is calculated. Both the nitrogen and phosphorus parts of the
nutrient component use enrichment ratios to estimate that portion of the two
nutrients transported with sediment. The pesticide component considers foliar
interception, degradation, and washoff, as well as adsorption, desorption, and
degradation in the soil. This method, like the nutrient model, uses enrichment
ratios and partitioning coefficients to calculate the separate sediment and wa-
ter phases of pesticide loss.

These models are compared in table 1-1. In addition to these models for
predicting runoff, erosion, and agricultural chemicals, several models were de-
veloped to estimate runoff and erosion for relatively large basins. Represen-
tative features are given in table 1-2.

Beasley, Monke, and Huggins (_1) developed a distributed deterministic mod-
el. (ANSWERS) for predicting runoff and erosion/sediment transport for different
agricultural management systems. The basic hydrologic component from Huggins
and Monke (j^) describes surface runoff, subsurface flow, and channel flow in a
system of square grids laid over the watershed. The infiltration element of
the model is basically the infiltration function of the USDAHL model (18) .
When the water content of the control zone exceeds field capacity, infiltrated
water becomes subsurface drainage. The erosion component of ANSWERS consists
of modifications of the USLE ( 330 • Two soil detachment processes were inclu-
ded: (a) rainfall detachment, described by Meyer and Wischmeier (22), and (b)
overland flow detachment, described by Foster (9_) . Sediment transport of both
overland and channel flow is based on transport capacity. Channel erosion is
assumed to be negligible, and only deposition is allowed in channel flow.

Simons, Li, and Ward (26) developed an event model to predict runoff and
sediment yield from small basins. The hydrologic component consists of the
kinematic wave model for overland flow and channel flow with infiltration ap-
proximated by the Green and Ampt (16) infiltration equation. The sediment com-
ponent considers erosion by raindrop splash and shear stress of overland flow.
Raindrop erosion is expressed as a power function of rainfall intensity and an
empirically determined erodibility factor. Erosion by overland flow uses a de-
tachment coefficient that requires calibration for specific soils. Sediment
transport in the model considers transport capacity for individual sediment
sizes. Bed load transport and suspended load transport are estimated.

Wade and Heady (29) developed an economic model based on agricultural crop
production considering sediment as a pollutant. The model, referred to as the

Table 1-1.— Water qual ity models, basic components and scale of application

Model

Date

Hydrology
component

Erosion/

sedimentation

component

Pesticide,,
component—

Nutrient ?,
component—

Scale of
application

PRT

1973

SWM

Negev

As,Ds,Vo,De

None

Field.

ACTMO

1975

USDAHL

Modified
USLE.

As,Ds,Vo,De

M,N,NL

Basin.

WASCH

1975

Parametric

None

Field.

ARM

1976

SWM

Negev

As.Ds.De

M,D,N,I,NL,AP,SP

Field.

NPS

1976

SWM

Negev

None

Basin.

Williams

1978

SCSCN

Williams-Modified
USLE.

None

M,D,N,I,NL,AP,SP

Basin.

RFF

1978

Mean river
flow(?).

Loading
functions.

Basin.

CREAMS

1979

SCSCN,
infiltration.

Interr ill-rill de-
tachment; overland
flow transport cap-
acity; concentrated
flow detachment and
transport capacity;
impoundment deposi-
tion.

As.Ds.Vo.De

M.N.NL.D,
AP.SP.

Field.

1/ No precedent for pesticide model; symbols for processes are:
As - adsorption; Ds - desorption; Vo - volatilization; De - degradation.

II No precedent for nutrient model; symbols for process are:
M - mineralization; D - denitrif ication; N - nitrification; I - imobil ization;
AP - adsorbed phosphorus; SP - solution phosphorus.

NL - nitrate leaching;

Table 1-2.— Hydrology-sedimentation models, basic components, and scale of application

Hydrology
component

Erosion
component

Scale

ANSWERS

1977

USDAHL infiltration;
kinematic flow; chan-
nel routing.

Interr i 1 1-r i 1 1 detach-
ment, overland and chan-
nel flow transport capa-
city.

Basin.

Simons and others 1977

Infiltration, kinema-
tic flow, channel
routing.

Raindrop and overland
flow detachment and
transport capacity,
channel flow detachment
and transport capacity.

Basin.

Wade and Heady 1978

Sediment delivery ratios,
sediment transport ratios.

Basin.

National Water Assessment (NWA) model, does not contain a hydrologic component
but estimates average annual erosion with the USLE (_33) for 105 Producing Areas
(PA) covering the United States. Sediment delivery ratios, estimated for each
PA by using measured and computed data, are used to estimate sediment delivery.
River basin sediment accounting is made by sediment ratios estimated for the
rivers of the PA's. River flow apparently is not used in the accounting sys-
tem, and the transport ratios are determined subjectively to give river sedi-
ment yields. Where lakes were involved in the river systems, estimated trap
efficiencies were used in determining transport ratios. Linear programming was
used with the NWA model to consider 5 sediment control alternatives to calcu-
late the associated sediment yield to the oceans from 18 river basins of the
United States.

REFERENCES

(1) Beasley, D. B., E. J. Monke, and L. F. Huggins.

1977. ANSWERS: A model for watershed planning. Purdue Agricultural
Experiment Station Journal Paper No. 7038. 34 pp.

(2) Bruce, R. R., L. A. Harper, R. A. Leonard, W. M. Snyder, and A. W.
Thomas.

1975. A model for runoff of pesticides from small upland watersheds.
Journal of Environmental Qual ity 4(4) :541-548.

(3) Crawford, N. H., and A. S. Donigian, Jr.

1973. Pesticide transport and runoff model for agricultural lands.
U.S. Environmental Protection Agency, EPA-660/2-74-013. 211 pp.
Washington, D.C.

(4) 1 an(j R. k. Linsley.

1962. The synthesis of continuous streamflow hydrographs on a digital
computer. Stanford University, Department of Civil Engineering,
Technical Report No. 12. Stanford, Calif.

(5) , and R. K. Linsley.

1966. Digital simulation in hydrology: Stanford Watershed Model IV.
Stanford University, Department of Civil Engineering, 210 pp. Stan-
ford, Calif.

(6) Davey, W. B.

1975. Conservation districts and 208 water quality management. U.S.
Environmental Protection Agency and National Association of Conserva-
tion Districts, 349 pp. U.S. Government Printing Office, Washington,
D.C.

(7) Donigian, A. S., Jr., and N. H. Crawford.

1976. Modeling pesticides and nutrients on agricultural lands. U.S.
Environmental Protection Agency, Environmental Protection Technology
Series, EPA-600/2-76-043, 317 pp. Washington, D.C.

(8) Donigian, A. S., Jr., and N. H. Crawford.

1976. Modeling nonpoint pollution from the land surface. U.S. Envi-
ronmental Protection Agency, Ecological Research Series, EPA-600/3-
76083, 279 pp. Washington, D.C.

(9) Foster, G. R.

1976. Sedimentation, general. Proceedings National Symposium on Urban
Hydrology, Hydraulics, and Sediment Control, University of Kentucky,
Lexington, Ken. July 26-29.

(10) , and L. D. Meyer.

1975. Mathematical simulation of upland erosion using fundamental ero-
sion mechanics. In: Present and prospective technology for predict-
ing sediment yield!- and sources. U.S. Department of Agriculture, Ag-
ricultural Research Servcice, Southern Region, ARS-S-40, pp. 190-207.
(Series discontinued; Agricultural Research Service is now Science
and Education Administration-Agricultural Research.)

(11) s L. D. Meyer, and C. A. Onstad.

1973. Erosion equation derived from modeling principles. Paper 73-
2550. Presented at 1973 Winter Meeting ASAE, December 11-14, 1973,
Chicago, 111.

(12) Frere, M. H., C. A. Onstad, and H. N. Holtan.

1975. ACTMO, an agricultural chemical transport model. U.S. Depart-
ment of Agriculture, Agricultural Research Service, Headquarters,
ARS-H-3, 54 pp. (Series discontinued; Agricultural Research Service
is now Science and Education Administration-Agricultural Research.)

(13) Frevert, R. K., G. 0. Schwab, T. W. Edminster, and K. K. Barnes.

1955. Soil and Water Conservation. John Wiley and Sons, Inc., New
York, N.Y. 479 pp.

(14) Gianessi, L. P., H. M. Peskin, and G. K. Young.

1978. A national water pollution network model. [Unpublished] 21 pp.

(15) Glymph, L. M., and H. N. Holtan.

1969. Land treatment in agricultural watershed hydrology research.
Effects of watershed changes on stream flow, Water Resources Sympos-
ium No. 2, University of Texas, Austin, Tex., pp. 44-68.

(16) Green, W. H., and G. A. Ampt.

1911. Studies on soil physics, Part I: The flow of air and water in
soils. Journal of Agricultural Science.

(17) Holtan, H. N., and N. C. Lopez.

1971. USDAHL-70 model of watershed hydrology. U.S. Department of Ag-
riculture, Technical Bulletin No. 1435, 84 pp. Washington, D.C.

(18) , G. J. Stiltner, W. H. Henson, and N. C. Lopez.

1975. USDAHL-74 revised model of watershed hydrology, a United States
contribution to the International Hydrological Decade. U.S. Depart-
ment of Agriculture, Technical Bulletin No. 1518, 99 pp. Washington,
D.C.

(19) Huggins, L. F., and E. J. Monke.

1966. The mathematical simulation of the hydrology of small water-
sheds. Purdue University, Water Resources Research Center, Technical
Report No. 1. 130 pp.

(20) James, L. D.

1970. An evaluation of relationships between streamflow patterns and
watershed characteristics through the use of OPSET: A self-calibrat-
ing version of the Stanford watershed model. University of Kentucky,
Water Resources Institute, Research Report No. 36. Lexington, Ken.

(21) McElroy, A. D., S. Y. Chiu, J. W. Nebgen, A. Aleti, and F. W. Bennett.

1976. Loading functions for assessment of water pollution from non-
point sources. U.S. Environmental Protection Agency, Environmental
Protection Technology Series, EPA-600/2-76-151, Washington, D.C. 444
pp.

(22) Meyer, L. D., and W. H. Wischmeier.

1969. Mathematical simulation of the processes of soil erosion by wa-
ter. Transactions of the American Society of Agricultural Engineers
12(6) :754-758.

(23) Negev, M. A.

1967 . Sediment model on a digital computer. Department of Civil En-
gineering, Stanford University, Technical Report No. 76, Stanford
Calif. 109 pp.

(24) Rawls, W. J., and L. E. Asmussen.

1973. Subsurface flow in the Georgia Coastal Plain. Journal of Irri-
gation and Drainage, Proceedings of the American Society of Civil En-
gineers Paper No. 10013, 99( IR3) :375-385.

(25) Ross, G. A.

1970. The Stanford watershed model: The correlation of parameter val-
ues selected by a computerized procedure with measurable physical
characteristics of the watershed. Water Resources Institute Research
Report No. 35. University of Kentucky, Lexington, 178 pp.

(26) Simons, D. B., R. M. Li, and T. J. Ward.

1977. A simple procedure for estimating on-site soil erosion. Pro-
ceedings of the International Symposium on Urban Hydrology, Hydrau-
lics, and Sediment Control, University of Kentucky, Lexington, July
18-21. pp. 95-102.

(27) Snyder, W. M.

1974. Development of a parametric hydrologic model useful for sediment
yield. In: Present and prospective technology for predicting sedi-
ment yieTds and sources. U.S. Department of Agriculture, Agricultur-
al Research Service, Southern Region, ARS-S-40. pp. 220-230. (Ser-
ies discontinued; Agricultural Research Service is now Science and
Education Administration-Agricultural Research.)

(28) Stewart, B. A., D. A. Woolhiser, W. H. Wischmeier, J. H. Caro, and M. H.
Frere.

1975. Control of water pollution from cropland. Vol. I - A manual for
guideline development. U.S. Department of of Agriculture, Agricul-
tural Research Service, Headquarters, Report No. ARS-H-5-1. Ill pp.
(Series discontinued; Agricultural Research Service is now Science
and Education Administration-Agricultural Research.)

(29) , D. A. Woolhiser, W. H. Wischmeier, J. H. Caro, and M. H. Frere.

1976. Control of water pollution from cropland. Vol. II: An over-
view. U.S. Department of Agriculture-Agricultural Research Service,
Headquarters, ARS-H-5-2. 187 pp. (Series discontinued; Agricultural
Research Service is now Science and Education Administration-Agricul-
tural Research.)

(30) U.S. Soil Conservation Service.

1972. SCS National Engineering Handbook, Sec. 4, Hydrology. 548 pp.

(31) Wade, J. C, and E. 0. Heady.

1978. Measurement of sediment control impacts on agriculture. Water
Resources Research 14(1) :l-8.

(32) Williams, J. R., and R. W. Hann, Jr.

1978. Optimal operation of large agricultural watersheds with water
quality constraints. Texas A&M University, Texas Water Resources In-
stitute, Technical Report No. 96, College Station. 152 pp.

(33) Wischmeier, W. H., and D. D. Smith.

1965. Predicting rainfall-erosion losses from cropland east of the
Rocky Mountains—Guide for selection of practices for soil and water
conservation. U.S. Department of Agriculture, Agriculture Handbook
No. 282. 47 pp.

(34) , and D. D. Smith.

1978. Predicting rainfall erosion losses--a guide to conservation
planning. U.S. Department of Agriculture, Agriculture Handbook No.
537. 58 pp.

Chapter 2. SIMULATION OF THE SURFACE WATER HYDROLOGY
R. E. Smith and J. R. Williams-^

INTRODUCTION

Central to the simulation of pollutant movement on and from a field site
is the simulation of the amount and rate of water movement on the surface and
through the soil. All major hydraulic processes which occur during a rainstorm

— such as rainfall infiltration, soil water movement, and surface water flow

— can be simulated in detail with current knowledge of hydraulics and the ca-
pabilities of modern computers. The constraint in the construction of this
model, however, is to approximate the complexity of these processes and their
interrelations with a model whose sophistication is appropriate to the detail
of data expected to be available in its intended use.

The field-scale hydrologic response simulation includes models for infil-
tration, soil water movement, and soil/plant evapotranspiration between storms.
It is a continuous simulation model using a day as the time step for evapora-
tion and soil water movement between storms, and using shorter time increments
dictated by available rainfall records during storms. The between-storm simu-
lation provides prediction of amount of seepage below the root zone and gives
an initial soil water content at the beginning of a storm, which is an impor-
tant initial condition for storm runoff simulation. When storm rainfall re-
cords are not available, runoff is estimated by the SCS curve number procedure
(Z).

INFILTRATION

Infiltration From Daily Rainfall
(SCS Curve Number Model)

The SCS curve number technique (_7) was selected for predicting runoff from
daily rainfall because (1) it is a familiar procedure that has been used for
many years in the United States; (2) it is computationally efficient; (3) the
required inputs are generally available; and (4) it relates runoff to soil
type, land use, and management practices. The use of readily available daily
rainfall is a particularly important attribute of the curve number technique.
For many locations, rainfall data with time increments of less than 1 day are
not available. Also, daily rainfall data manipulation and runoff computation
are more efficient than similar operations with shorter time increments.

1/ Hydraulic engineer, USDA-SEA-AR, Fort Collins, Colo., and hydraulic en-
gineer, USDA-SEA-AR, Grassland-Soil and Water Research Laboratory, Temple,
Tex., respectively.

Traditionally, the SCS has used an antecedent rainfall index to estimate
antecedent moisture as one of three conditions (I - dry, II - normal, and III -
wet). The relation between rainfall and runoff for these three conditions is
expressed as a curve number (CN). Each storm in a rainfall series is assigned
one of the three curve numbers according to antecedent rainfall. In reality,
CN varies continuously with soil moisture, and thus has many values instead of
only three. Runoff prediction accuracy was increased by using a soil moisture
accounting procedure to estimate the curve number for each storm (9_) . Although
the soil moisture accounting model was found to be superior to the antecedent
rainfall method, it did not contain a percolation component or a physically
based water balance. Also, the model required calibration with measured runoff
data.

Here the curve number technique was linked with evapotranspiration and
percolation models to form a model capable of maintaining a continuous water
balance. Calibration is not necessary, because the new model is more physical-
ly based. Besides predicting daily runoff volumes, an equation was also devel-
oped for predicting peak runoff rates. Tests with data from watersheds in
Texas, Nebraska, Georgia, Ohio, Oklahoma, Arizona, New Mexico, West Virginia,
Mississippi, Iowa, and Montana indicate that the model simulates runoff volumes
and peak rates realistically (tables 1-3 and 1-4).

Model Description

Runoff is predicted for daily rainfall using the SCS equation

q = (P - °-2s)2 r-Tin

g P + 0.8s Li iJ

where Q is the daily runoff; P is the daily rainfall; and s is a retention pa-
rameter, all having dimensions of length. The retention parameter s is related
to soil water content with the equation

s = s Vm^JM^ [i_2]

mx [ UL

where SM is the soil water content in the root zone, UL is upper limit of soil
water storage in the root zone, and smx is the maximum value of s. The maximum
value of s is estimated with the I moisture condition CN using the SCS (7_)
equation

[1-3]

where CNj is the moisture condition I CN. An estimate of the moisture condi-
tion II CN can be obtained easily for any watershed using the SCS Hydrology
Handbook (7_) . The corresponding CNj values are also tabulated. For computing
purposes CNj was related to CN j i with the polynomial

CNj = -16.91 + 1.348(CNn) - 0.01379(CNn )2 + 0.0001177(CNn )3. [1-4]

If soil water is distributed uniformly in the soil profile, equation [1-2]
should give a good estimate of the retention parameter, and thus the runoff.
However, if the soil water content is greater near the surface, equation [1-2]
would tend to give low runoff predictions. Conversely, runoff would be over-
predicted if the soil water content was greater in the lower root zone. To ac-
count for the soil water distribution, a weighting technique was developed.
The root zone was divided into seven layers and weighting factors (decreasing
with depth) were applied. The depth-weighted retention parameter is computed
with the equation

s = sr

1.0

i=l

SM-j

[1-5]

where W-j is the weighting factor, SM-j is the water content, and UL-j is the up-
per limit of water storage in storage i. J The weighting factors decrease with
depth according to the equation

Wi = 1.016

(5iVI

.16\RD/|

[1-6]

where D-j is the depth to the bottom of storage i, and RD is the root zone

N
depth. Equation [1-6] assures that ^Wi = 1.

1*1

The evapotranspiration and percolation components of the model are de-
scribed below. Since the model maintains a continuous water balance, mixed
land use watersheds are subdivided to reflect differences in ET for various
crops. Thus, runoff is predicted separately for each subarea and combined to
obtain the total runoff for the watershed. Division by land use increases ac-
curacy and gives a much better physical description of the water balance.

Peak runoff rate is predicted with the equation

= 200(DA)0-7(CS)0-159 (Q) (0.917DA0-0166) (lw)-0.187

[1-7]

3 2

where qp is the peak runoff rate in ft /s; DA is the drainage area in mi ; CS

is the mainstem channel slope in ft/mi; Q is the daily runoff volume in in; and
LW is the length-width ratio of the watershed. Data from 304 storms that oc-
curred on 56 watersheds located in 14 states were used to develop equation [I-
7]. Watershed areas ranged from 0.275 to 24 mi'2. Since these areas are larger
than what is usually considered field-scale, the equation has variable expo-
nents for DA and Q to accommodate areas down to 1 acre or less. These variable
exponents simply prevent unreasonably high predictions for small areas.

Model Testing and Evaluation

The runoff model based on the SCS curve number technique has been tested
on basins in Texas, Ohio, Georgia, Oklahoma, Nebraska, Arizona, New Mexico,
West Virginia, Mississippi, Iowa, and Montana. Results of the tests are shown

in tables 1-3 through 1-6. Table 1-3 shows that the model generally approxi-
mates long-term water yield (average annual runoff) well. Also, average ET and
percolation predictions seem realistic. The monthly R2 values shown in table
1-3 were obtained by comparing measured and predicted monthly runoff. Table
1-4 contains statistics obtained by comparing measured and predicted individual
runoff events. Although some of the R2 values are lower than desirable, the
standard deviations of the measured and predicted runoff are similar. This in-
dicates that the model simulates runoff with a frequency distribution similar
to that of the measured runoff, although the measured record is not duplicated
precisely. There are many reasons for prediction errors. Some more important
reasons are (1) the curve number system's inability to consider rainfall inten-
sity, duration, or distribution; (2) the use of average values for temperature,
solar radiation, and leaf area index instead of actual values; (3) lack of in-
formation on planting and tillage dates and incomplete soils descriptions; and
(4) errors in rainfall and runoff data.

Table 1-5 shows a comparison of measured and predicted percolation for wa-
tershed Z at Tifton, Ga. The measured values are actually subsurface flow mea-
sured at the watershed outlet. Of course, the predicted percolation is the
amount of water that flows downward below the root zone. Considering these
differences, the test can only indicate that the percolation model gave reason-
able results.

Table 1-6 contains measured and predicted percolation and evapotranspi ra-
tion for watershed 115 and lysimeter Y103A near Coshocton, Ohio. The measured
values were obtained from the lysimeter. Both the watershed and the lysimeter
had the same crop each year. Close comparisons between measured and predicted
values indicate satisfactory test results. Limited data prohibit percolation
and ET model tests as extensive as those of the runoff model.

Infiltration Simulation from Breakpoint Rainfall Data

Whenever rainfall information is available in terms of actual time pattern
of rainfall intensity or rate, the present understanding of soil water dynamics
allows a significantly improved prediction of infiltration and runoff as com-
pared with predictions based on amount of rain alone, such as the SCS curve
number method discussed above.

Infiltration during rainfall is composed of two phases, as illustrated in
figure 1-3. At the beginning of a rain, the soil has an initial saturation not
necessarily uniform with depth, but here assumed to be uniform in the (usually)
small upper region which most affects infiltration. The saturation, S-j , is de-
fined as

Si - £ [1-8]

where e-j is initial water content by volume, and<pis porosity.

In the early stages of rainfall, the surface saturation increases from S ] to a
maximum value S0 (theoretically, S0 —1), if the rainfall lasts long enough.
For S >_ S0, the soil controls surface flux, and the time when this begins is

Table 1-3. — Runoff model test results (annual-monthly)

Length

Average

Annual

Drainage
area

of
record

measured
P

predi

icted

Percolation

Monthly

Watershed location

(mi2)

(*r)

in)

(jn)

(in)

SW-2.

. Riesel . Tex.

0.004

36.94

6.28

7.73

27.82

1.14

0.75

SW-12. Do.

.005

38.08

9.05

6.45

30.63

1.05

.72

Y-6,

Do.

.025

38.08

5.72

7.34

30.10

.80

.86

Y-8,

Do.

.032

.08

6.72

6.14

31.22

.68

.65

21-H,

, Hastings, Nebr.

.006

.83

3.40

3.69

19.16

.03

.41

3-H,

Do.

.006

23.25

5.22

5.31

18.03

.01

.66

3-H,

Do.

.006

.45

4.75

5.41

17.98

.01

.68

P-li

Watkinsville, Ga.

.010

47,

.54

8.72

8.30

33.03

6.95

.46

P-2,

Do.

.005

44.26

5.94

6.46

29.15

9.30

.53

104,

Coshocton, Ohio

.002

38.04

.35

.61

32.39

4.75

.39

104,

Do.

.002

.40

.88

1.14

28.67

4.73

.92

129,

Do.

.004

35.73

.83

.84

29.81

5.15

.33

130,

Do.

.003

.50

.95

.84

30.58

4.18

.45

132,

Do.

.001

.32

2.08

2.18

28.70

4.57

.51

115,

Do.

.003

.07

1.93

2.33

31.22

3.53

.56

110,

Do.

.002

35.38

1.70

1.78

30.81

2.76

.43

118,

Do.

.003

.53

2.01

2.23

30.95

3.35

.53

106,

Do.

.002

.60

2.06

1.73

30.21

2.63

.33

192,

Do.

.012

.71

2.61

1.88

30.19

2.87

.48

R-5,

Chickasha, Okla.

.037

.14

1.76

1.95

27.31

o70

.73

R-7,

Do.

.030

.14

5.98

5.30

24.19

.41

.86

C-4,

Do.

.047

.21

3.45

2.79

29.33

.10

.59

C-5,

Do.

.020

.46

2.02

1.92

25.32

.09

.35

W-6,

Cherokee, Okla.

.003

.74

3.33

3.58

20.05

.17

.45

W-7,

Do.

.003

.74

3.59

20.03

.17

.53

W-13

, Do.

.003

.79

1.66

2.11

19.56

.02

.59

W-2,

Guthrie, Okla.

.005

28.00

4.18

3.74

23.34

1.17

.85

W-l,

Do.

.004

.41

.67

.89

25.54

1.57

.46

W-2,

Vega, Tex.

.150

18.54

.97

.80

17.91

.00

.27

W-l,

Spur, Tex.

.018

20.07

1.93

2.05

18.03

.00

.67

W-2,

Do.

.015

20.07

2.68

2.56

17.51

.00

.70

W-3,

Do.

.018

.10

1.55

1.72

18.35

.00

.72

63105, Lucky Hills, Ariz.

.001

,15

1.14

1.00

10.32

.00

.84

01, 1

■"t. Stanton, New Mex.

.038

14.40

.02

.05

14.96

.00

.24

02,

Do.

.050

14.64

.00

.06

14.94

.00

.001

66001, Moorefield, W. Va.

.013

.16

2.90

2.77

25.95

.93

.51

62014, Holly Spr., Miss.

.002

45.46

15.08

15.98

27.95

1.82

.80

62015, Do.

.002

.73

9.48

9.65

24.22

1.19

.65

22003, Guthrie Ctr., Iowa

.019

24.31

1.16

1.10

22.59

.28

.74

Z, Tifton, Ga.

.001

50.65

2.96

3.04

41.25

7.17

.26

A, Sidney, Mont.

.003

.50

1.70

1.32

13.65

.00

.72

W-3,

Garland, Tex.

.016

.02

9.14

8.92

30.66

.86

.84

W-l,

Do.

.039

.24

5.11

6.42

31.68

3.50

.86

W-3,

Tyler, Tex.

.012

.35

1.31

1.79

31.93

8.20

.36

W-5,

Do.

.003

.56

8.23

7.25

30.90

3.52

.58

W-4,

Do.

.093

.03

7.63

6.90

30.32

3.52

.60

Table 1-4. — Runoff model test results (events)

Runoff volume
Standard deviation
Measured Predicted

Peak

runoff rate

Mean

Standard
Measured

deviation

Watershed location

Measured

Predicted

SW-2, Riesel, Tex.

0.85

0.74

2.16

1.74

3.12

2.39

SW-12, Do.

.69

.74

.55

1.72

1.46

2.53

2.34

Y-6, Do.

.90

.68

.84

5.34

5.36

7.48

8.78

Y-8, Do.

.64

.50

5.90

5.76

8.76

8.29

21-H, Hastings, Nebr.

.46

.42

.37

1.88

1.37

2.66

2.23

3-H, Do.

.65

.47

.41

3.06

1.51

4.05

2.45

3-H, Do.

.55

.56

3.06

2.96

4.05

4.64

P-l, Watkinsville, Ga.

.60

.61

.48

4.47

3.01

6.72

4.94

P-2, Do.

.64

.45

.44

1.88

1.48

2.63

2.49

104, Coshocton, Ohio

.28

.10

.11

.48

.28

.75

.43

104, Do.

.88

.42

.36

.48

.45

.75

1.09

129, Do.

.24

.25

.17

.57

.67

1.05

1.30

130, Do.

.29

.26

.16

.33

.74

.58

132, Do.

.46

.33

.24

.08

.09

.11

.16

115, Do.

.55

.32

.29

.73

.57

1.31

1.16

110, Do.

.37

.32

.23

.34

.45

.82

.86

118, Do.

.52

.29

.23

.59

.60

1.10

1.02

106, Do.

.31

.22

.19

.54

.49

1.15

.87

192, Do.

.41

.37

.25

1.05

1.24

2.88

2.41

R-5, Chickasha, Okla.

.72

.35

.32

5.69

5.39

9.99

8.15

R-7, Do.

.86

.45

.44

7.65

6.59

11.50

10.81

C-4, Do.

.64

.42

.36

3.75

1.69

3.74

2.75

C-5, Do.

.46

.32

.29

1.45

1.07

1.48

2.00

W-6, Cherokee, Okla.

.35

.44

.41

1.32

1.48

1.68

2.28

W-7, Do.

.42

.49

.42

1.48

1.35

1.87

2.09

W-13, Do.

.59

.31

.33

1.33

1.21

1.74

1.95

W-2, Guthrie, Okla.

.67

.38

.36

1.65

1.61

2.45

2.09

W-I, Do.

.15

.22

.15

.44

.75

.55

.89

63105, Lucky Hills, Ariz.

.64

.23

.17

.29

.13

.46

.20

01, Ft. Stanton, New Mex.

.003

.02

.12

1.43

1.84

1.03

2.23

02, Do.

.10

.00

.13

1.00

1.49

.36

2.49

66001, Moorefield, W. Va.

.71

.70

1.54

.48

.64

.44

1.11

62014, Holly Spr., Miss.

.82

.74

.64

.89

1.39

1.21

1.65

62015, Do.

.62

.57

.50

.89

1.00

1.21

1.39

22003, Guthrie Ctr., Iowa

.44

.16

.17

.63

.95

.28

1.28

Z, Tifton, Ga.

.08

.23

.48

.61

.54

.77

A, Sidney, Mont.

.68

.34

.28

.48

.38

.88

.59

Table 1-5. — Percolation model results

at Tifton,

Ga.} watershed

Annual

Average

monthly

percolat
Measured

ion (in)
Predicted

Month

percolat

.ion

(in)

Year

Measured

Predicted

1970

17.90

19.74

2.15

2.67

1971

9.23

15.77

2.95

2.67

1972

11.72

12.13

1.52

1.58

1973

17.42

12.78

2.54

1.89

1974

8.41

10.60

.78

1.03

1975

14.83

9.81

.57

.71

Mean

13.25

13.47

.40

.13

Std. deviation

4.09

3.70

Mean

.98
.71
.00
.00
.72
1.11

.72
.24
.42
.25
1.16
1.12

Std

. deviation

.97

.90

/rdt = P

/-r(t)

/fdt = F

TpA —

\ f=fp+f'

y Ilk /~f(t)

■■■■l

tp TIME

Figure 1-3. — Definition diagram for infiltration model

O HCOLD

■^

O^HCVJCOLf)Lf)Ln«a-CMt-l

LO ID N CO Oi O rH CSJ <OT3
,_, ,_, H CD

■*^lOOOO

O co CO r-. en

<Tt VD Lf> <d" vo ro

CT> l£> LT> CM CO CT> CM

iD^or^oOLDoor^mvoo

3 S

S X _

O +J O O -MOO

T3 C fl-CD CIDT3T3
<0t-CU<0<0S-<D<T3(0

0) o r a) o

3 s: s:

+-> o o +-> o o +->

(DT3-0 C W-OT3 C TO
CD <1> O -C CU CD

32:z:o3s:s:o3s:2:o3

^■unor-.oocnorHMM<tanoNcocrio-H(\j c xj
^j-^-^-»d-«a-«d-Lf)Lf>LnLnLf)LOur>Lf)Lf>Lr>(£>^D<£) <o

CT> CTt C7> CX> CTt CTt CT» CT* CX> CX» CX> CX* CX> CT* CT> CX> CTt CX> CTt CD •

4->
OO

O *d"

o «*

J-

<T>

(/>

5- -O

1/1

<T3

s».

.-<

-a

called time of ponding, t

The dept21of rain which enters the soil prior

to tD is analogous to the SCS CN parameter
Unlilce Ig, F is a function of rainfall rate, r.

and is here called F,

Rainfall infiltration models must describe both the occurrence of tn and

the shape of the subsequent infiltration curve,
storms will have short total time so that tr _< t.

f(t), (t > tp). Clearly, some

The rainfall model based on soil water flow theory employed here is relat-
ed to an infiltration curve describing infiltration from an instantaneous pond-
ing condition. It is important to understand that the sudden ponding condition
is distinct from the condition where water arrives at the soil surface at a
certain rate, as for rainfall, yet the infiltration curves are mathematically
related (5).

Others could be chosen, but this model employs the Green and Ampt U) in-
filtration relation

Kst = F -

c o

S^ln

1 +

*Hc<So " V

[1-9]

in which Ks = effective saturated conductivity, [LT~ ]

t = time from start of ponding, [T]

Hc = effective capillary tension, a soil parameter, [L], and

F = cumulative depth of infiltration, [L].

Derivation of this expression may be found elsewhere (1_, 5J . This relation
will be used in different forms below to obtain expressions for ponding time
and the inter-storm infiltration-rate curve.

Ponding Time

The key to predicting ponding time is a relation of infiltrated depth, in-
filtration rate, and time. The relationship used is derived from experiments
and theory which indicate that

tp t(rp)
Fp =) r(t)dt = J f0(t)dt

Cl-10]

in which r is rainfall rate, rp is rainfall rate at tp. In this equation,
f0 is the infiltration rate curve for the instantaneous ponding condition
(r = <*>), which equals dF/dt, with F(t) defined by equation [1-9] above. In
other words, for the rainfall rate, r„, the depth of water, Fn, infiltrated

which f0 = rp

equal to the depth of water1" infiltrated from t = 0
for the case of infiltration from sudden ponding.

.o the time at

The Green and Ampt [I) model comes originally from an assumption of a pis-
ton-type movement of soil water downward in the soil, as in figure 1-4. This
model can be used with equation [1-10] to derive an expression for ponding
time. Clearly, by continuity,

SOIL WATER POTENTIAL, H SOIL WATER CONTENT, 6

1 t

Figure 1-4. — Conceptual assumptions in the Green-Ampt (JJ infiltration model.

F = L*(S0 - Si)

Cl-ll]

where L is the depth of wetting. Total head across the wetting front is -(Hc+
L), and by Darcy's law, infiltration rate, f, is

f = K

Hc + L

^s\ L
Solving equation [1-11] for L, and substituting into equation [I-12J

f = K.

This can be solved for F(f), as

HC«P(S0-S.) +F

"c^o' VKs
(f - Ks)

[1-12]

[1-13]

[1-14]

For a ponding time expression, equation [1-10] can be written as

Fp = / fdt = F(rp)

so that ponding depth is estimated as

C v 0

S,)KS

(rp - Ks>

GDK(

LI-15J

in which G is used for Hc, and D represents <t>(S0 - S-j)

This expression is used to find when ponding is expected to occur in a
histogram of rainfall rate pulses of variable height (as with breakpoint rain-
fall data). If ponding occurs within a pulse [t-j < tp < t-j + ]_], interpolation
is used.

Infiltration Curve

To obtain an expression for infiltration within a breakpoint interval

>n [I-( ~

G = H,

where t > tp, we start with equation [1-9], and as above for simplicity let

and

so that

Si)*

D = (S,

Kst-F-GD inl +-

[1-16]

This expression may be derived from equation [1-13] by setting f = dF/dt and
integrating. We assume a finite difference perturbation of equation [1-16] to
be equally correct, so that

Ks(t + At) = F + AF - GD lnjl +

F + AF
GD

[1-17]

Subtracting equation [1-16] from equation [1-17] and rearranging the logarith-
mic expression, we form the difference expression

KsAt= AF- GDlnfl+^L

[1-18]

Then using the first term of the following series approximation for the natural
logarithm,

ln(l + x) = 2
we can solve for AF as

2+x 3 \ 2 + x

+ .

AF = /2K At(GD + F) + (F - j<S At)2 -. (F - <S At'
5 2 2

with an approximation error of approximately 8% (3).

[1-19]

Let A = KsAt/2, so that

AF = y4A[GD + F] + (F - A)2 + A - F . [1-20]

Mean infiltration rate can be calculated for time interval Ati as

- AFi

fi ■ ATT • CI-21]

Runoff during interval i is q^ = rs - f ^ . Calculation proceeds through
the storm, with F being updated at each interval i as

F-j+i = Fi + AFi [1-22]

in any interval where t > tp. Where ri At < At AFi from equation [1-18], then

Fi+i = Fi + HAt . [1-23]

Total runoff for a storm having n intervals is simply

n
Q =2qiAti. [1-24]

i = l

Adjustments for Hourly Data

On the basis of rainfall record analysis, it has been found (2_) that storm
intensity changes significantly during intervals shorter than 60 minutes, and
that peak intensity will be significantly biased (reduced) by hourly data.
Therefore, employment of hourly data such as are commonly available through the
National Weather Service (formerly U.S. Weather Bureau) suggests an adjustment
of the infiltration procedure. Sufficient research has not been completed to
know an optimum adjustment, nor how the adjustment should be changed to reflect
various climatic zones. In this model, the procedure adopted in the interim is
to base predicted ponding times for hourly data on 133% of the hourly intensity
for the early storm periods, still using equation [1-15]. Storm EI (see volume
III, chapter 1 for definition) is calculated using a 30-minute maximum inten-
sity which is assumed to be twice the maximum hourly intensity.

Multiple Storms

More than one storm is assumed to occur on a day when a rainfall hietus of
te = 180 minutes is found. In this case, D? = S0 - Si is estimated for the sub-
quent storm as

D2 =<P[0.9 - (te/180)0.05]. [1-25]

This estimates a rather wet initial condition for the next storm, and Q2 for
the second storm is added to Qi for the first to get the daily Q. The same
process can be used if more than two storms occur.

Estimating Runoff Peak Rates

The breakpoint rainfall infiltration simulation produces a histogram of
excess rainfall rates which are sequences of time intervals and associated
rainfall excess rates. These rates are typically much larger than that seen as
the output runoff rate from a field or small watershed. To estimate peak run-
off rates from field areas, we can use kinematic surface water flow equations.
The description here is brief, and a more detailed explanation may be obtained
by referring to Woolhiser (10).

Runoff begins when free surface water is generated from the excess of
rainfall rate above the infiltration rate. A certain "lag" period must occur,
however, in which rate of rainfall excess far exceeds the runoff rate at the
catchment outlet. If rainfall excess rate is uniform and lasts long enough,
"equilibrium" will eventually occur when the two rates are equal.

Since rainfall excess usually varies rather abruptly during a storm and
rarely lasts long enough, equilibrium flow is practically a hypothetical con-
cept. Shallow water flow hydraulics allows estimation, nevertheless, of peak
runoff rates.

Figure 1-5 illustrates some of the aspects of the flow regimes occurring
during surface water flow response to rainfall. This figure presents, graphic-
ally, the movement of the characteristic "waves" which originate from the

TIME

STEADY, NONUNIFORM
FLOW REGION

3 2

RAINFALL EXCESS, i

DISTANCE ALONG SURFACE. X

Figure 1-5. — Illustration of flow regimes occurring
during serf ace-flow response to rainfall.

upstream point (x = 0), and illustrates the usual case where the peak rainfall
excess occurs prior to the "equilibrium" time. A hypothetical pattern of rain-
fall excess rates is illustrated on the left of the ordinate.

The curve from x = 0, t = 0 is cal led the upstream characteristic. To the
right of this line, flow is unsteady (rising), but uniform. To the left of
this characteristic, flow is steady but nonuniform, varying with location along
the surface.

Along any characteristic, between points (2) and (3) for example, the
depth is

hj " hj-l + 1j*tj CI-26]

where ij is rainfall excess rate during interval j, At is time from start of
interval j, and hj_i is depth on the characteristic at time tj_i- The velocity
of the characteristics is

v = mahm_1 = dx/dt [1-27]

where v = wave celerity (not flow velocity)

m = uniform flow exponent (1.5 for the Chezy roughness law)

a = uniform flow coefficient (= C/S for the Chezy roughness law)

S = plane slope, and C = Chezy roughness coefficient.

Combining equations [1-26] and [1-27] and integrating for the characteristic
starting at t; i, we have

f (hM + '* jst)m"lds +

x. =ma / ihj-1 +ijStj ds + x._.

= f (hj._1 + ijAt) + Xj_j . [1-28]

This gives us the position of the "wave" from the upstream edge at any time.
This equation may be applied successively with all increments of i(t) to obtain
the distance the peak (or any other disturbance) moves. In any case, the peak
runoff rate may be estimated from the greatest depth h reached at x = L accord-
ing to equation [1-26] for the fastest characteristic since at all points for
kinematic flow,

q = a h m . [1-29]

This estimation procedure is based on the peak flow occurring with mono-
tonic rise of depth, since if flow recession occurs prior to a second rainfall
burst, recession calculations are necessary. In a very few cases, this could
cause underestimation of the flow peak.

To estimate the peak outflow from a complex rainfall pattern, we must
choose the characteristic path along which (in time) the largest rates of rain-
fall excess occur, and thus, the largest depth h at the downstream edge of the
surface or watershed outlet. Obviously, this characteristic would usually in-
clude the time in which the largest rainfall excess occurs, plus the intervals

with largest i before and after, necessary for the characteristic to traverse
the distance L (figure 1-5). One may estimate the peak, therefore, by looking
at the characteristic incremental depth h as given in equation [1-26] for the
peak excess interval, and adding adjacent intervals (of largest positive excess
rate) to each side of the peak, thus choosing the fastest characteristic for
which x from equation [1-28] equals or exceeds the length L.

In estimating peak runoff rates using kinematic surface water flow equa-
tions, complex slopes can be represented by hydraulical ly equivalent uniform
slopes. If the length is broken into N regions of different slope and rough-
ness, as illustrated in figure 1-6, the equivalent single plane values can be
determined. For each segment or sub-plane j, where j = 1, N, there is an aj as
in equation [1-27]

*j ■ CJ

[1-30]

RAINFALL EXCESS, i

I J J I I i I I I 1 i

C-

Figure 1-6. — Representation of a complex slope
in terms of a single equivalent plane.

For the Manning roughness law, C is the same as 1.49/n where n is Manning's
roughness coefficient. The objective is to determine the ac for a single
plane which best represents the composite hydraulic response of the set of N
planes. From equation [1-30], if Sc is the overall slope as in figure 1-6,
then composite ac will specify a Cc and vice versa.

Research by Wu (11) indicates that the best hydraul ically equivalent
single plane is the one that gives equilibrium surface detention storage equal
to that of the set of different planes. Detention storage, A, (10) is

hdx

[1-31]

where h is local surface water depth. Thus equation [1-31] is the equivalence
criteria. Equilibrium discharge for any plane of length x is, by continuity,

Li iter ia. ui|U i I iui i uiti uii^nai yc iui any pianc ui iciiyuii a

ix, where i is rainfall excess. From equation [1-29], then

ix = a h(x)m
and from which equilibrium depth is

h.U|

1/m

[1-32]

[1-33]

Equating storage on the equivalent single plane to that on the set of planes

n
we have A = V' A j . This combined with equations [1-31] and [1-32], yields

C 1/m fl 1/m f2 . 1/m /N .

J(fc) **') (?) **y (?) -**••;/ (?

0 ° Xl Vl

1/m

b(«c

/m b
L

1 / i \1/(T1 v b /i \1/m/ b b\ /i \1/[T1 h

tfcr) x +fe) (x* -xi>-ft) (L -

(N-1

where b = (m + l)/m. Dividing both sides by (-H (i) /m (L) gives

1/m

1/m/ b

and rearranging

C /S~ = ci =
c V c c

where x0 = 0 and xpj = L

Z, b I
x. - x . .
_J 1=1
ojl/ra

j=l

[1-34]

[1-35]

[1-36]

Evaluation Results

Table 1-7 presents results of model testing performed on those watersheds
where breakpoint rainfall and runoff records were obtained, including two wa-
tersheds where lysimeter data provided estimates of amounts of seepage below

c <o C CUCM
C <L> 3 > S_

<C B S- <U

<NJ CTi

CO CO

§ c

-z I

r-t <n

»-> co cr>

c O C\J

^H CNJ

(O E CD

r- .r- >,

■r- in

-O "O •■-

o s-

•r- CD
+-> -O

+-> D. Q_

Ol 3 .—

d. a. •
ai a) en

CD CD CNJ

(/Il/IH

■— l CvJ CD CO

the root zone. Complicating the ability to accurately predict runoff on agri-
cultural watersheds, as mentioned above, is the occurrence of often unrecorded
cultivation practices which severely modify the soil's infiltration properties.
This is demonstrated in the relative predictive accuracy for the cultivated wa-
tershed at Tifton, Ga., and the rangeland at Lucky Hills, Ariz. In addition,
the data include many instances of errors in time, such as major storms re-
corded by only two breakpoints, or records of runoff attributed to days where
the major portion of the rainfall was a day earlier. Other common data errors
include blank periods where major storms occur with no recorded runoff, and
runoff peaks with greater rates than the associated rainfall rate. Curiously,
most of the examples in this table show correlation coefficients for daily run-
off in the 0.80 to 0.90 range, yet with occasional years having contrastingly
low r2 of 0.1 to 0.2.

EVAP0TRANSPIRATI0N AND SOIL WATER ROUTING

From either infiltration submodel, water that enters the soil, F, becomes
either evapotranspiration, storage, or seepage below the root zone. A daily
time interval is used between storm events, and the components of the water
balance equation are evaluated. In equation form,

SMi = SMi-i + Fj - ET-j - 0i + Mi [1-37]

where Fi = infiltration on day i

ET-j = plant and soil evapotranspiration on day i

0-j = seepage below the root zone on day i

M-j = snowmelt amount on day i

SM = soil water storage in the root zone.

Snowmelt

A simple snow accumulation and snowmelt equation is used by the model
taken from Stewart and others (£) . For all those days where precipitation oc-
curs when the temperature is less than 0°C, that precipitation is stored in the
form of snow. When snow storage exists and the temperature, T, is above 0° C,
snowmelt occurs, and input to the soil at the surface is calculated by

M. = 0.18T [1-38]

unless M is greater than the amount of surface snow. Although this model is
quite simplistic, it does help account for spring melt input, and would be dif-
ficult to improve without detailed daily temperature and radiation information.

Evapotranspiration

As illustrated in figure 1-7, the soil water balance model considers both
soil and plant evaporation losses, and treats the growth of plant leaf area and
depth of root extraction explicitly. The evapotranspiration (ET) component of
the runoff model is taken from Ritchie (4). To compute potential evaporation,
the model uses the equation

1.28 A H(
E° A + Y

[1-39]

where E0 is the potential evaporation; A is the slope of the saturation vapor
pressure curve at the mean air temperature; H0 is the net solar radiation; and
y is a psychrometric constant. A is computed with the equation

. _ 5304 (21.255 - 5304/T)
T2*

where T is the daily temperature in degrees kelvin. H0 is calculated with the
equation

= (1 - X)(R)
ho 58.3

[1-41]

where R is the daily solar radiation in langleys and X is the albedo for solar
radiation.

/Jit

I t t j fcRAIN OR SNOW
* I I k

INFILTRATION I

v.-. .•■•:<>:;• infiltration sensitive

'^t-' ' '■ — j— SOIL DEPTH

MAXIMUM

ROOTING

OEPTH

Figure 1-7. — Schematic representation of the
water-balance model.

Soil Evaporation

The model computes soil and plant evaporation separately. Potential daily
soil evaporation is predicted with the equation

r c "0.4 LAI rT .ol

Eso = E0 e [1-42]

where Eso is the potential evaporation at the soil surface and LAI is the leaf
area index defined as the area of plant leaves relative to the soil surface
area. Actual soil evaporation is computed in two stages. In the first stage,
soil evaporation is limited only by the energy available at the surface, and
thus is equal to the potential soil evaporation. When the accumulated soil
evaporation exceeds the stage one upper limit, the stage two evaporative pro-
cess begins. Here the stage one upper limit is estimated with the equation

U - 9 (os - 3)0,42 [1-43]

where U is the stage one upper limit in mm and as is a soil evaporation parame-
ter dependent on soil water transmission characteristics (ranges from about 3.3
to 5.5 mm/d1/^). Ritchie (4_) suggests a value of 4.5 for loamy soils, 3.5 for
clays, and 3.3 for sands.

Stage-two daily soil evaporation is predicted with the equation

Es - as [t1/2 - (t - 1)1/2] [1-44]

where Es is the soil evaporation for day t, and t is the number of days since
stage two evaporation began.

Plant Transpiration
Plant evaporation is computed with the equations

Ep = UoHLAU s 0 < LAI < 3 [1-45]

Ep = E0 - Es , LAI > 3 . [1-46]

If soil moisture is limited, plant evaporation is reduced with the equation

(ED)(SM)

Ep, = P 1 , SM < 0.25FC [1-47]

KL 0.25FC

where Ep is normal plant evaporation: Ep|_ is plant evaporation reduced by limi-
ted SM, and FC is the field capacity of the soil. Evapotranspiration, the sum
of plant and soil evaporation, cannot exceed E0.

Drought

When soil moisture falls below 15 bar amount (estimated), plant growth is
stopped by holding leaf area index constant until water becomes available.

This allows an interaction between rainfall data and leaf area index descrip-
tion, to account in an approximate manner for drought conditions.

PERCOLATION

The model uses a soil storage routing technique to predict flow through
the root zone (8). When the SCS (7_) curve number method is used, the root zone
is divided into seven layers or storages for routing. Root-zone depth is usual-
ly estimated to be three feet, although it may vary with various crops and
soils. The routing equation is

°=nF + §)• (f + fi)> fc [i-48]

where F is the infiltration or inflow rate; ST is the storage volume; a is the
storage coefficient; and At is the routing interval (1 day). If inflow plus
storage does not exceed field capacity, FC, percolation is not predicted to oc-
cur. The storage coefficient is a function of the travel time through the
storage expressed by the equation

• ■ wht CI-49]

where t is the travel time through a storage. Travel time is estimated with
the equation

t - ^V^ [1-50]

where SM is soil water storage, and rc is the saturated conductivity of the
soil .

Besides percolation losses, each soil storage is subject to ET losses.
Therefore, the daily predicted ET must be distributed properly through the
storages. A model for simulating root growth is used for this purpose. The
water-use rate as a function of root depth is expressed by the equation

-4.16RD rT r-,-,

u = u0 e [1-51]

where u is the water-use rate by the crop at depth, RD, and u0 is the rate at
the surface. The total water use within any depth can be computed by integra-
ting equation [1-51] to obtain

ct - uo m -4.16RD* rT co-,

ET = ^-jg (1 - e ) . [1-52]

The value of u0 is determined for the root depth each day, and the water use in
each storage is computed with the equation

^Oe^'4-16™1-1-6"4'16"01) "-5«

where uwj is the water use in storage, i, and RD-j_i and RD-j are the depths at
the top and bottom of storage, i.

When the breakpoint infiltration model is used for runoff calculations,
the soil water movement and percolation calculation involves only two storage
elements, a surface soil zone, and a root soil zone. The surface soil zone is
subject to soil evaporation from the evapotranspiration model, plus a portion
of the plant root extraction. It is the region of the soil which determines
initial conditions to which the infiltration model is sensitive. The lower
zone is subject to root extraction during the growing season. A root growth
model is used in this option which simulates relative root depth proportional
to relative leaf area index.

Water moves from the upper soil zone to the root zone as a function of the
positive difference in saturation between the two zones as:

qs = Cs SS3(SS - Sp)*Ds, (Ss > Sp), [1-54]

in which qs = daily water movement from surface to root zone
Cs = coefficient (normally 0.1)
Ss = saturation by volume in surface zone
Sp = saturation by volume in root zone
<p = porosity
Ds = depth of surface zone (2 to 5 cm)

This is designed as a crude analogy to Darcy's law, with CSSS approximat-
ing the relation between conductivity and saturation.

Seepage from the root zone is predicted to occur when Sp exceeds field ca-
pacity, and is estimated as the daily excess of Ss over field capacity. Root
extraction occurs from both surface and root zones in proportion to the rela-
tive root depth, which varies with leaf area index up to the maximum depth.
Thus, if root depth = 2D , evapotranspiration water is taken equally from Ds
and root zone, D. Total soil water storage UL is estimated as porosity times
surface depth, D , plus field capacity in the root zone. Field capacity is a
ratio, Fc, of porosity, so that

UL = <PDS + fc • Dp . [1-55]

REFERENCES

(1) Green, W. A., and G. A. Ampt.

1911. Studies on soil physics, I. The flow of air and water thru
soils. Journal of Agricultural Science, 4:1-24.

(2) Hershfield, D. M.

1961. Rainfall frequency atlas of the United States. Weather Bureau
Technical Paper No. 40, 115 pp. (Weather Bureau is now the National
Weather Service.)

(3) Li, R. M., M. A. Stevens, and D. B. Simons.

1976. Solutions to Green-Ampt infiltration equation. Journal of Irri-
gation and Drainage Division, American Society of Civil Engineers, 102
(IR2):239-248.

(4) Ritchie, J. T.

1972. A model for predicting evaporation from a row crop with incom-
plete cover. Water Resources Research 8( 5) : 1204-1213 .

(5) Smith, R. E., and J. Y. Parlange.

1978. A parameter-efficient hydrologic infiltration model. Water Re-
sources Research 14(3) :533-538.

(6) Stewart, B. A., and others.

1975. Control of water pollution from cropland. Vol. 1, A manual for
guideline development. U. S. Department of Agriculture, Agricultural
Research Service, Headquarters, ARS-H-5-1. (Series discontinued; Ag-
ricultural Research Service now Science and Education Administration-
Agricultural Research.)

(7) U.S. Department of Agriculture, Soil Conservation Service.

1972. National engineering handbook, Hydrology, Section 4, 1972. 548
pp.

(8) Williams, J. R., and R. W. Hann.

1978. Optimal operation of large agricultural watersheds with water
quality constraints. Texas A&M University, Texas Water Resources
Institute, TR-96. 152 pp.

(9) , and W. V. LaSeur.

1976. Water yield model using SCS curve numbers. Journal of the Hy-
draulics Division, American Society of Civil Engineers 102(HY9):
1241-1253.

(10) Woolhiser, D. A.

1974. Unsteady free-surface flow problems. Ch. 12, Proceeding of the
Institute on Unsteady Flow in Open Channels, Colorado State Univer-
sity, July 17-28, Water Resources Publications, Ft. Collins, Colo.

(11) Wu, Y. H.

1978. Effects of roughness and its spatial variability on runoff hydro-
graphs. PhD Dissertation, Colorado State University, Civil Engineer-
ing Department Report No. CED77-78 YHW7, Fort Collins, Colo., 174 pp.

Chapter 3. A MODEL TO ESTIMATE SEDIMENT YIELD FROM FIELD-SIZED AREAS:

DEVELOPMENT OF MODEL

G. R. Foster, L. J. Lane, J. D. Nowlin, J. M. Laflen, and R. A. Young!/

INTRODUCTION

Estimates of erosion and sediment yield on field-sized areas are needed to
wisely select best management practices to control erosion for maintenance of
soil productivity and to control sediment yield for prevention of excessive
degradation of water quality. A field is a typical management unit for farm-
ers. The selection of a management practice is usually based on site-specific
conditions. Soil conservationists have used the Universal Soil Loss Equation
(USLE) (31) for several years to select practices specifically tailored to a
given farmer's situation. Assuming that sediment yield tolerance for mainten-
ance of water quality will be established for given local areas, best manage-
ment practices can then be selected based on a given farmer's needs and the
tolerable water loading for fields in his area using a model such as the one
described herein (8) .

Sediment yield is a function of detachment of soil particles and the sub-
sequent transport of these particles (sediment). On a given field, either
detachment or sediment transport capacity may limit sediment yield depending on
topography, soil characteristics, cover, and rainfall /runoff rates and amounts.
Control of sediment yield by detachment or transport can change from season to
season, from storm to storm, and even within a storm. The relationship for
detachment is different from the one for transport so that they cannot be lump-
ed together into a single equation. Since detachment and transport for each
storm are best considered separately, lumped equations such as the USLE (an
erosion equation), or Williams' (29) modified USLE (a flow transport, sediment
yield equation) cannot give the best results over a broad range of conditions
on field-sized areas. Furthermore, the interrelation between detachment and
transport is nonlinear and interactive for each storm, which prevents using
separate equations to linearly accumulate amount of detached sediment or sedi-
ment transport capacity over several storms. Therefore, to simulate erosion
and sediment yield on an individual storm basis and to satisfy the need for a
continuous simulation model, a rather fundamental approach was selected where
separate equations are used for soil detachment and sediment transport.

A number of fundamentally based models (_1, 20) compute detachment and
transport at various times during the runoff event. While these models are
powerful, their excessive use of computer time practically prohibits simulating
20 to 30 years of record. The model described herein uses characteristic

1/ Hydraulic engineer, USDA-SEA-AR, Lafayette, Ind.; hydrologist, USDA-
SEA-AR, Tucson, Ariz.; computer programmer, Agricultural Engineering Depart-
ment, Purdue University, Lafayette, Ind.; agricultural engineer, USDA-SEA-AR,
Ames, Iowa; and agricultural engineer, USDA-SEA-AR, Morris, Minn.

rainfall and runoff factors for a storm to compute detachment and sediment
transport for that storm. In terms of computational time, this amounts to a
single time step for models which simulate over the entire runoff event.

The model is intended to be useful without calibration or collection of
research data to determine parameter values. Therefore, established relation-
ships such as the USLE were modified and used in the model.

OVERVIEW OF THE MODEL

Every model is a representation and a simplification of the prototype.
Various techniques, including planes and channels (20), square grids (_1) , con-
verging sections (28), and stream tubes (24) have been used. Most erosion/se-
diment yield models have adequate degrees of freedom to fit observed data.
Some models, depending on their representation scheme, distort parameter values
more than do others. Distortion of parameter values greatly reduces the trans-
ferability of parameter values from one area to another (18). An objective in
this model development was to represent the field in a way that minimizes para-
meter distortion. Hydrologic input to the erosion/sediment yield component
consists of rainfall amount, rainfall erosivity (EI), runoff volume, and peak
rate of runoff. These terms drive soil detachment and subsequent transport in
overland and open channel flow.

Overland flow, channel flow, and impoundment (pond) elements are used to
represent major features of a field. The user selects the best combination of
elements and enters the appropriate sequence number according to table 1-8.
The model (computer program) calls the elements in the proper sequence. Typi-
cal systems that the model can represent are illustrated in figure 1-8.

Table 1-8. — Possible elements and their calling sequence used to represent

field-sized area

Sequence number Elements and their sequence

1 Overland

2 Overland-Pond

3 Overland-Channel

4 Overland-Channel-Channel

5 Overland-Channel-Pond

6 Overland-Channel-Channel-Pond

Computations begin in the uppermost element, which is always the overland
flow element, and proceed downstream. Sediment concentration (for each parti-
cle type) is the output from each element which becomes the input to the next
element in the sequence.

BASIC CONCEPTS

Basic Equations

Sediment load is assumed to be limited by either the amount of sediment
made available by detachment or by transport capacity (11). Also, quasi steady

state is assumed so that a rainfall and a runoff rate characteristic of each
storm can be used in the computations. Sediment movement downslope obeys con-
tinuity of mass expressed by:

DL + DF

[1-56]

where qs = sediment load per unit width per unit time, x = distance, D|_ = lat-
eral inflow of sediment (mass/unit area/unit time), and Dp = detachment or dep-
osition by flow (mass/unit area/unit time). The assumption of quasisteady
state allows deletion of time terms from equation [1-56]. The major sequence
of computations is illustrated in figure 1-9.

OVERLAND FLOW
SLOPE REPRESENTATION

OVERLAND FLOW

*<ev

AVERAGE SLOPE

4 •'4' _ _

(I) OVERLAND FLOW
SEQUENCE AND SLOPE REPRESENTATION

IMPOUNDMENT
TERRACE

CONCENTRATED FLOW

(2) OVERLAND FLOW
POND SEQUENCE

(3) OVERLAND FLOW
CHANNEL SEQUENCE

OVERLAND FLOW

CHANNEL FLOW

OUTLET
CHANNEL FLOW

(4) OVERLAND FLOW
CHANNEL-CHANNEL SEQUENCE

(5) OVERLAND FLOW
CHANNEL-POND SEQUENCE

Figure 1-8. — Schematic representation of typical field
systems in the field-scale erosion/sediment yield
model .

SEDIMENT LOAD

FROM UPSLOPE

SEGMENT

COMPUTE SEDIMENT
ADDITION FROM
LATERAL INFLOW

SUM SFDIMENT

LOAOS FOR AN

INITIAL-POTENTIAL

SEDIMENT LOAD

COMPUTE TRANSPORT

CAPACITY BASED ON

POTENTIAL SEDIMENT

LOAD

COMPUTE SEDIMENT

LOAD LEAVING

SEGMENT

COMPUTE NEW POTENTIAL

SEDIMENT LOAD AS

SUM OF SEDIMENT FROM

DETACHMENT CAPACITY

AND INITIAL-POTENTIAL

SEDIMENT LOAD

COMPUTE TRANSPORT

CAPACITY BASED ON

NEW POTENTIAL

SEDIMENT LOAD

SEDIMENT LOAD LEAVING
SEGMENT EQUALS
NEW POTENTIAL
SEDIMENT LOAD

IT FLOW DETACHMENT
0 THAT WHICH WILL
JST FILL TRANSPORT

SEDIMENT LOAD

LEAVING -EGMENT

EQUALS TRANSPORT

CAPACITY

Figure 1-9. — Flow chart for detachment-transport-deposition
computations within a segment of overland flow or con-
centrated flow elements.

Lateral sediment inflow is from intern* 11 erosion on overland flow ele-
ments, or it is from overland flow (or a channel, if two channel segments are
in the sequence) for the channel elements. Flow in rills on overland flow
areas or in channels transports the sediment load downstream. Lateral sediment
inflow is assumed regardless of whether the flow is detaching or depositing.

For each segment, either on the overland flow element or in a channel, the
model computes an initial potential sediment load which is the sum of the sedi-
ment load from the immediate upslope segment plus that added by lateral inflow
within the segment. If this potential load is less than the flow's transport
capacity, detachment occurs at the lesser of the detachment capacity rate or
the rate which will just fill transport capacity. When detachment by flow
occurs, it adds particles, having the particle-size distribution for detached

sediment given as input. No sorting is allowed during detachment.

If the initial potential sediment load is greater than the transport
capacity, deposition is assumed to occur at the rate of:

D = a (Tc - qs) [1-57]

where D = deposition rate (mass/unit area/unit time), a = a first order reac-
tion coefficient (length-1), and Tc = transport capacity (mass/unit width/unit
time). The coefficient a is estimated from:

a= e^x [1-58]

where e = 0.5 for overland flow ( 5J , and 1.0 for channel flow ( 7J , Vs= particle
fall velocity, and qLx = qw = discharge per unit width (volume/unit width/unit
time). Fall velocity is estimated assuming standard drag relationships for a
sphere of a given diameter and density falling in still water.

Detachment-Deposition Limiting Cases

Four possible cases may exist for a segment: (1) Deposition may occur
over the entire segment; (2) detachment by flow in the upper end and deposition
in the lower end may (but not necessarily) occur when transport capacity de-
creases in a segment; (3) deposition on the upper end and detachment by flow in
the lower end may (but not necessarily) occur when transport capacity increases
within the segment; (4) detachment by flow may occur all along the segment.

Case 1 occurs when Tc < qs all along the segment. Where deposition occurs
over the entire segment length, deposition rate is:

D = [*/(l+*)](dTc/dx - DL) [l - (xu/x)l+*j + Du(xu/x)l+* [1-59]

where

♦ = £Vs/qL [1-60]

where dTc/dx is assumed constant over the segment and Du = deposition rate at
xu. The deposition rate Du may be estimated from:

Du = « (TCu " qsu) L"I-61]

where Tcu and qsu= respectively, the transport capacity and sediment load at
xu. Sediment load at x is:

qs = Tc - D/o [1-62]

Case 2 occurs when Tcu> qsu, dTc/dx < 0, and Tc becomes less than qs with-
in the segment. If dTc/dx < 0 for a segment where Tcu > qsu, Tc may decrease
below qs within the segment. The point where qs = Tc is determined as x^.
This becomes xu in equation [1-60], with Du = 0. Deposition and sediment load
are computed from equations [1-59], [1-60], and [1-62].

Case 3 occurs when Tcu < qsu, dTc/dx > 0, and Tc becomes greater than qs
within the segment. In situations like a grass buffer strip, the transport
capacity at the upper edge may drop abruptly to a level below the sediment
load. Within the upper end of the strip, the sediment load decreases due to
deposition while the transport capacity increases from the point of the abrupt
decrease. Somewhere upslope from the lower edge of the strip, the sediment
load equals the transport capacity. At this point, x^e » deposition ends, that
is, Du = 0, and Tc = qs. Downslope, detachment by flow occurs. The point where
deposition ends is given by:

where:

xu (l - C(l+*)/*][Du/(dTc/dx - DL)]\ 1/(1+*} [1-63]
Du ■ o (Tcu " qsu) [1-64]

and Tcu = transport capacity after the abrupt decrease at xu and qsu = sediment
load at xu. Continuity of sediment load is maintained, but D may be discontin-
uous at segment ends.

Downslope from xde , where detachment by flow occurs, the sediment load is
given by:

qs = (DFu + DLu + DFL + DLL)Ax/2 + gsu [1-65]

where the second subscript u or L indicates upper or lower, and Ax = length of
the segment where detachment by flow is occurring. In this case, ax is from
X(je to the lower end of the segment; qsu is at x<je, which is Tc at x<-|e, Dpu = 0
at x^es and Dpi_ is either detachment capacity at x or that which will just fill
the transport capacity.

Case 4 occurs when Tc > qsu over the entire segment. Sediment load is
computed with equation [1-65].

The equation for sediment transport capacity, Tc , shifts total transport
capacity among the various particle types. If transport capacity exceeds avail-
ability for one particle type while it is less for another, transport capacity
is shifted from the particle type having the excess to the one having the defi-
cit. Furthermore, logic checks within the model prevent simultaneous deposi-
tion and detachment of particles by flow.

Eroded sediment is a mixture of particles having various sizes and densi-
ties. The distribution is broken into classes, with each class represented by
a particle diameter and density. Equations [1-58] to [1-65] are solved for
each particle type within the constraints noted above.

SEDIMENT CHARACTERISTICS

Sediment eroded on field-sized areas is a mixture of primary particles and
aggregates (conglomerates of primary particles). The distribution of these
particles as they are detached is a function of soil properties, management,
and rainfall and runoff characteristics. If deposition occurs, usually the
coarse and dense particles are deposited first, leaving a finer sediment mix-
ture. The input to the model is the distribution of the sediment as it is
detached; the model calculates a new distribution if it calculates that
deposition occurs.

Based on our survey of existing data, values given in table 1-9 are an ex-
ample of input for many midwestern silt loam soils.

Table 1-9. — Sediment characteristics assumed for detached sediment before depo-
sition; assumed typical of many midwestern silt loam soils

Particle
type

Diameter

Specific
gravity

Fraction of total

amount

(mass basis)

Primary clay
Primary silt
Small aggregate
Large aggregate
Primary sand

(mm)
0.002
.010
.030
.500
.200

(q/cm3:
2.60
2.65
1.80
1.60
2.65

0.05
.08
.50
.31
.06

If the particle distribution is not known, the model assumes five particle
types, and estimates the distribution from the primary particle-size distribu-
tion of the soil mass from the following equations:

PSA =

(1.0 - 0RCL

PSI = 0.13 0RSI

PCL =

3.2 0RCL

2 0RCL

SAG = '

0.28(0RCL

k 0.57

LAG =

L.O - PSA -

,2.49

0RSA

0.25) + 0.5

PSI - PCL - SAG

0RCL

< 0.25

0.25

< 0RCL

0.5 <

' 0RCL

0.50

[1-66]

[1-67]

[1-68]

[1-69]
[1-70]
[1-71]

[1-72]

if LAG < 0.0, multiply all others by the same ratio to make LAG = 0.0 where
0RCL, 0RSI, and 0RSA are, respectively, fractions for primary clay, silt, and
sand in the original soil mass, and PCL, PSI, PSA, SAG, and LAG are, respec-
tively, fractions for primary clay, silt, sand, and small and large aggregates
in the detached sediment.

The diameters for the particles are given by:
DPCL = 0.002 mm
DPSI = 0.010 mm
DPSA = 0.20 mm

[1-73]
[1-74]
[1-75]

0.03 mm ORCL < 0.25 [1-76]

\ 0.20(0RCL - 0.25) + 0.03 mm 0.25 < ORCL ^0.60 [1-77]

0.1 mm 0.60 < ORCL [1-78]

DLAG = 2(0RCL) mm [1-79]

where DPCL, DPSI, DPSA, DSAG, and DLAG are, respectively, the diameters of the
primary clay, silt, and sand, and the small and large aggregates in sediment.
The assumed specific gravities are shown in table 1-9. The primary particle
composition of the sediment load is estimated from:

Small aggregates:

CLSAG = SAG • ORCL/ (ORCL + 0RSI) [1-80]

SISAG = SAG • 0RSI/(0RCL + ORSI) [1-81]

SASAG = 0.0 [1-82]

Large aggregates:

CLLAG = ORCL - PCL - CLSAG [1-83]

SILAG = ORSI - PSI - SISAG [1-84]

SALAG = ORSA - PSA [1-85]

where CLSAG, SISAG, and SASAG = fractions of the total for the sediment of, re-
spectively, primary clay, silt, and sand in the small aggregates in the sedi-
ment load, and CLLAG, SILAG, and SALAG are corresponding fractions for the
large aggregates.

If the clay in the large aggregate expressed as a fraction for that parti-
cle alone is less than 0.5 times ORCL, the distribution of the particle types
is recomputed so that this constraint can be met. A sum, SUMPRI, is computed
whereby:

SUMPRI = PCL + PSI + PSA. [1-86]

The fractions PSA, PSI, and PCL are not changed. The new SAG is:

SAG = (0.3 + 0.5 SUMPRI)(ORCL + ORSI )/[l - 0.5 (ORCL + ORSI)]. [1-87]

Equation [1-87] is derived given previously determined values for PCL, PSI, and
PSA; the sum of primary clay fractions for the total sediment equals the clay
fraction in the original soil, and the assumption that the fraction of primary
clay in LAG equals half of the primary clay in the original soil.

The model also computes an enrichment ratio using specific surface areas
for organic matter, clay, silt, and sand. Organic matter is distributed among
the particle types based on the proportion of primary clay in each type. En-
richment ratio is the ratio of the total specific surface area for the sediment
to that for the original soil.

Although these relationships are approximations to the data found in the
literature (33), they represent the general trends.

OVERLAND FLOW ELEMENT

Detachment Equation

Detachment on intern' 11 and rill areas and transport and deposition by
rill flow are the erosion-transport processes on the overland flow element.
Detachment is described by a modified USLE (21) written as:

DLi = 0.210 EI (s + 0.014) KCP (ap/Vu) [1-88]

and

DFr = 37983 mVuap1/3 (x/72.6)"1"1 s2 KCP (ap/Vu) [1-89]

where Dm = intern 11 detachment rate (lb/ft^/s), Dpr = rill detachment capaci-
ty rate (lb/ft^/s), EI = Wischmeier's rainfall erosivity (energy times 30-min-
ute intensity) [100(ft-tons/acre)(i n/hr)], x = distance downslope (ft), s =
sine of slope angle, m = slope length exponent, K = USLE soil erodibility fac-
tor [(tons/acre)(acre/100 ft-tons)(hr/i n)], C = soil loss ratio of the USLE
cover-management factor, P = USLE contouring factor, Vu = runoff volume [vol-
ume/unit area (ft)], and ap = peak runoff rate [volume/unit area/unit time (ft/
/s)]. Note that for P only the contouring factor of the USLE is used. The
model is structured to directly account for other USLE P-f actor effects such as
strip cropping and deposition in terrace channels. These P factors are highly
variable, and the model can account for a number of them.

Storm Erosivity

The hydrologic processes of rainfall and runoff drive the erosion-trans-
port processes. Storm EI (storm energy times maximum 30-mi nute intensity),
volume of runoff, and peak discharge are the variables used to characterize hy-
drologic inputs. Values for these factors are generated by the hydrology com-
ponent of CREAMS. When daily rainfall amounts are used, storm EI is estimated
from (21):

EI = 8.0 VR1,51 [1-90]

where EI = storm EI [(100 ft-tons/acre)(i n/hr)] and VR = volume of rainfall
(in). This equation is very approximate. It was developed by regression anal-
ysis from about 2,700 data points used in the development of the USLE and has a
coefficient of determination (R2) of 0.56. When breakpoint rainfall is used,
storm EI is computed using standard USLE procedures (3JJ . Storm energy per
unit of rainfall is given by:

,10, [1-91]

where e = rainfall energy per unit of rainfall (ft-tons/acre-i n) and i = rain-
fall intensity (in/hr). The energy for each segment of the rainfall hyetograph
is the product of e and the rainfall amount for the segment. Total energy for

the storm is the sum of these incremental energies.

Slope Length Exponent, m

For slopes less than 150 ft, m is set to 2.0, but for slopes longer than
150 ft, m is limited by:

m = 1.0 + 5.011/ln (x). [1-92]

This limit avoids excessive erosion for very long slopes (12). Equation [1-92]
limits the effective slope length exponent for the total of rill and interrill
erosion to 1.67 so far as it is a function of length. The effective exponent
is a function of slope (smaller for flatter slopes), runoff erosivity relative
to rainfall erosivity (greater for relatively greater runoff erosivity), and
slope length (greater with longer length, except with the above restriction).

The Yalin sediment transport equation (32) is used to describe sediment
transport capacity. It gave reasonable results when compared with experimental
data for deposition of sand and coal by overland flow in a laboratory study (5_,
9.) and on field plots (vol. Ill, ch. 10). The Yalin equation was modified to
distribute transport capacity among the various particle types. The discussion
of the method given below is abstracted from Foster and Meyer (10), Davis (5),
and Khaleel and others (14).

The Yalin equation is given by

(Sg)gpw dV,

= 0.635 <5

[l - 1 In (1 +o)] = Ps [1-93]

where:

a = A • 6 [1-94]

6 = ~~ - 1 (when Y < Y . 6 = 0) [1-95]

' rr Cr

2.45(Sgy°'4(Ycr)1/2 [1-96]

Y = (Sq - 1.0)gd tI-97]

V* = (gRSf)1/2 [1-98]

where V* = shear velocity =(t/pw)1/2,t = shear stress, g = acceleration of gra-
vity, pw = mass density of the fluid, R = hydraulic radius, Sf = slope of the
energy gradeline, Sg = particle specific gravity, d = particle diameter, Ycr =
critical lift force given by the Shields' diagram extended to low particle
Reynolds numbers (22), and Ws = transport capacity (mass/unit width/unit time).
The constant 0.635 and Shields' diagram were empirically derived.

The sediment load may have fewer46articles of a given type than the flow's
transport capacity for that type. At the same time, the sediment load of other
particle types may exceed the flow's transport capacity for those types. The
excess transport capacity for the deficit types is assumed to be available to
increase the transport capacity for the types where available sediment exceeds
transport capacity.

The Yalin equation was modified to shift excess transport capacity. For
large sediment loads (sediment loads for each particle type clearly in excess
of the respective transport capacity for each particle type), or for small
loads (sediment loads for each particle type clearly less than the respective
transport capacity for each particle type), the flow's transport capacity is
distributed among the available particle types based on particle size and den-
sity and flow hydraulics (10) .

Yalin assumed that the number of particles in transport is proportional to
6. For a mixture, the number of particles of a given type i is assumed to be
proportional to 6-j . Values of -j for each particle type in a mixture are cal-
culated and summed to give a total:

i = l

[1-99]

where ns = number of particle types. The number of transported particles of
type i in a mixture is given as:

(Ne)i = Ni (6-j/T) [1-100]

where N^ = number of particles transported in sediment of uniform type i for a

As derived by Yalin, the nondimensional transport, Ps, of equation [1-93]
is proportional to the number of particles in transport. Then

(PJ, = Vi CI-101]

e l T

where (Pe)i = the effective Ps for particle type i in a mixture, and P -j is the
Ps calculated for uniform material of type i. The transport capacity WS1- of
each particle type in a mixture is then expressed by:

Wsi ■ <Vi 'Vi^V*- [I"102]

This is the transport capacity assuming that the supply of all particle types
is either greater than or less than their respective Ws-j. When availability of
some types is greater than their Wsi and others are less than their WS1- , trans-
port capacity shifts from those types where supply is less than capacity so
that all of the total transport capacity is used.

The steps given below are followed to redistribute the transport capacity
when excesses and deficits occur.

1. For those particles where WS1- ^ qsi (qsi = sediment load for particle
type i), compute the actual required P-jreq from equation [1-93], that

pireq = qsi/(sg) igpwdiv*« [1-103]

2. For those particle types where WS1- >_qsi» the sum:

SPT = X (pireq/Pi) [1-104]

i-1

is computed where k-j = 1 if Ws-j >. qsi ^nd k-j = 0 if WS1- < qsl- . The

sum SPT represents the fraction of the total transport capacity used

by those particle types where sediment availability is less than
transport capacity.

3. The excess (expressed as a fraction of the total) to be distributed

Exc = 1 - SPT . [1-105]

4. For those particle types where WS1- < qsl-, sum 6 i as:

SDLT =X<5 i1!' [1-106]

i=l

where 1-j = 0 if Ws-j >^ qsj and 1 -j = 1 if WS1- < qsi«

5. The excess is distributed according to the distribution of 6-j among
the particle types, that is,

Tci = (6i/SDLT)(Exc)(Pi)(Sg)igpwdiV^li , [1-107]

and

Tci = qsi M . H-108]

6. Repeat steps 1-5 until either all T c-j < qsl- or all T C1- >_ qs-j . When
the former occurs, the proper Tcj's have been found. If the latter
occurs, one particle type will have all of the excess transport capac-
ity. The excess for this one type should be equally distributed among
all of the types. This is done by:

SMUS = X(pireq/pi) [1-109]

i=l

Tci = (1.0/SMUS) qsi . [1-110]

Conversion from Storm to Rate Basis
Without the (ap/Vu) term, equations [1-88] and [1-89], as originally

developed (12), were on a storm basis, while the transport equation is on an
instantaneous rate basis. The two are combined by assuming that the computed
sediment concentration represents an average for the runoff event, and that the
peak discharge represents a characteristic discharge that can be used to com-
pute the average concentration.

Since most field-sized areas are relatively small, time of concentration
is usually small and is assumed to be less than rainfall duration. Thus, for a
given storm, discharge at a location is assumed to be directly proportional to
upstream drainage area.

Shear Stress

The transport equation requires an estimate of shear stress. The sediment
transport concept, where shear is divided between form roughness and grain
roughness, is used to estimate the shear stress acting on the soil, the portion
assumed responsible for sediment transport (13). Mulch or vegetation reduces
this stress. The shear stress acting on the soil, tS01--|, is estimated by:

T soil = Yys(nbov/ncov)°'9 [1-111]

where Y = weight density of water, y = flow depth for bare, smooth soil, s =
sine of slope angle, nbov = Manning's n for bare soil (0.01 assumed), and ncov
= total Manning's n for rough surfaces or soil covered by mulch or vegetation.
Flow depth is estimated by the Manning equation as:

Y - [qwnbov/s1/2]°*6 [I-H2]

where qw = rate of discharge per unit width. Although the Darcy-Weisbach equa-
tion with a varying friction factor for laminar flow might be more accurate for
y in some cases, most users are better acquainted with estimating Manning's n.
The error in estimating a value for the roughness factor is probably greater
than the error in using the Manning equation for laminar flow.

Slope Segments

Computations begin at the upper end of the slope. Sediment is routed
downslope much the same as it is in most erosion models. The output is the
sediment concentration for each particle type. Concentration multiplied by the
runoff volume and overland flow area represented by the overland flow profile
gives the sediment yield for the storm on the overland flow area of the field.

The overland flow area is represented by a typical land profile selected
from several possible overland flow paths. Its shape may be uniform, convex,
concave, or a combination of these shapes. Inputs are total slope length,
average steepness, steepness at the upper end, steepness at the lower end, and
location of the end points of a mid-uniform section.

CX, ,Y,)

COORDINATES OF POINTS
A, C, AND D AND SLOPES S,
AND S GIVEN AS INPUT

(X4,Y4)

Given this minimum of in-
formation, the model establishes
segments along the profile. The
procedure is illustrated by the
convex shape shown in figure I-
10. Coordinates of points A, C,
and D are given, as are slopes
S5 and Sm. A quadratic curve
will pass through point C tan-
gent to line CD and through
point E tangent to line AB. The
location of point E is the in-
tersection of a line having a
slope equal to the average of S^
and Sm with line AB. If X2
is less than X ^ , X3 is shifted
downslope so that Xj = X2.

DISTANCE

Figure 1-10. — Representation of convex
slope profile for overland flow.

Each uniform section is one
segment. In figure I -10, AE and
CD are segments. Convex sections
like EC are divided into only
three segments, because detach-
ment and transport computations are not especially sensitive to the number of
segments on convex slopes. Concave segments are divided into 10 segments be-
cause deposition computations on concave slopes are especially sensitive to the
number of segments. Furthermore, several segments are required to accurately
determine where deposition begins.

Additional segment ends are designated where K, C, P, or n change. Given
locations where these changes occur as input, the model computes the coordi-
nates for all the segments for the overland flow slope.

Selection of Parameter Values

Slope length is, perhaps, the most difficult of the overland flow parame-
ters to estimate. Williams and Berndt's (30) contour method is a possible
technique to use. Another is to sketch flow lines from the watershed boundary
to concentrated flow. Topography in most fields causes overland flow to con-
verge into concentrated flow within about 300 ft. Certainly a grass waterway
or a terrace channel is the end of overland slope length.

Values for the parameters K, C, and P (contouring) are selected from
Wischmeier and Smith (31) according to crop stage. Values for Manning s ncov
may be selected from Lane and others (18) or from vol. II, ch. 2.

CHANNEL ELEMENT

The channel element is used to represent flow in terrace channels, diver-
sions, major flow concentrations where topography has caused overland flow to
converge, grass waterways, row middles or graded rows, tail ditches, and other
similar channels. The channel element does not describe gully or large stream

channel erosion.

With the exception that shear stress and detachment by flow are estimated
differently, the same concepts and equations are used in both the channel and
overland flow elements. Discharge along the channel is assumed to vary direct-
ly with upstream drainage area. A discharge greater than zero is permitted at
the upper end to account for upland contributing areas. As with the overland
flow element, changes in the controlling variables along the channel are allow-
ed. Thus, a channel with a decreasing slope or a change in cover can be ana-
lyzed.

Spatially Varied Flow Equations

Flow in most channels in fields is spatially varied, especially for out-
lets restricted by ridges and heavy vegetation, and for very flat terrace chan-
nels. Also, discharge generally increases along the channel. The model ap-
proximates the energy gradeline along the channel using a set of normalized
curves and assuming steady flow at peak discharge. As an alternative, the mod-
el will set the friction slope equal to the channel slope.

The equation for spatially varied flow (3_) with increasing discharge in a
triangular channel may be normalized as:

dy 2 16/3 4 2 5

a£ ■ [S, - C2 Xyy4 - C3 IJy, Ml - C3 X+ /y# ] [1-113]

where y* = y/ye, y = flow depth, ye = flow depth at the end of the channel, S*

= S • L ff/y , X = distance along channel, X* = X/L ff, and L_eff = effective

channel length (that is, the length of the channel if it is extended upslope to

where discharge would be zero with the given lateral inflow rate). Constants
C]_ , C2, and C3 are given by:

Ci = [z5/2/2(z2 + 1)1/2 ]2/3 n_114]

C2 = We n Leff1/2/Ciye19/6]2 f>115^

C3 ■ 2 3 Qe2/g z2 ye5 [1-116]

where n = Manning's n, z = side slope of channel, Qe = discharge at end of
channel, 3 = energy coefficient [1.56 used from McCool and others (23)], and g
= acceleration due to gravity. Equation [1-113] was solved for a range of typ-
ical values of (4 , C2, and C3. The curves given by equations [1-117] to [1-126]
were fitted by regression to the solutions.

Range of C3: C3 > 0.3

where 0.0 < S* < 1.2 and X* < 0.9,

SSF = 0.2777 - 3.3110 X* + 9.1683 X*2 - 8.9551X*3 [1-117]

where 1 .2 < S* < 4.8 and X* < 0.9, [1-117]

SSF = 2.6002 - 8.0678X* + 15.6502X*2 - 11.7998X*3, [1-118]

where 4.8 < S* < 20.0 and X* < 0.9,

SSF = 3.8532 - 12.9501X* + 21.1788X*2 - 12.1143X*3, [1-119]

and where 20.0 < S* and X* < 0.9»

SSF = 0. [1-120]

Range of C3: 0.3 > C3 > 0.03

Where S* > 0 and X* < 0.8,

SSF = 2.0553 - 6.9875X* + 11.418X*2 - 6.4588X*3, [1-121]

and where S* = 0 and X* < 0.9,

SSF = 0.0392 - 0.4774X* + 1.0775X*2 - 1.3694X*3. [1-122]

Range of C3: 0.03 > C3 > 0.007

Where S* > 0 and X* < 0.8,

SSF = 1.5386 - 5.2042X* + 8.4477X*2 - 4.740X *3, [1-123]

and where S* = 0.0 and X* <^ 0.9,

SSF = 0.0014 - 0.0162X* - 0.0926X 2 - 0.0377X 3. [1-124]

Range of C3: 0.007 > C3

Where S* > 0 and X* < 0.7,

SSF = 1.2742 - 4.7020X* + 8.4755X*2 - 5.3332X*3, [1-125]

and where S* = 0 and X* _< 0.9,

SSF = - 0.0363X*2. [1-126]

With these values of SSF, the friction slope is:

Sf - (S* - SSF) ye/Leff . [1-127]

Flow depth ye at the end of the channel is estimated by assuming at the
user's option, either critical depth, depth of uniform flow in an outlet con-
trol channel, or depth from a rating curve.

A triangular channel section, a reasonable approximation to most field
channels, was used to develop the friction slope curves because the equations
are simpler. In the model, a triangular channel must be used to estimate the
slope of the energy gradeline, but the user may select a triangular, rectanqu-
lar, or "naturally eroded" section in other computational components of the
channel element.

Concentrated Flow Detachment

In the spring after planting, concentrated flow from intense rains on a
freshly prepared seedbed often erodes through the finely tilled layer to the
depth of secondary tillage or, perhaps, primary tillage. Once the channel
erodes to the nonerodible layer, it widens at a decreasing rate.

Data from observed rill erosion rates (vol. Ill, ch. 11) suggests that de-
tachment capacity (lb/ft2/s) by flow over a loosely tilled seed-seedbed may be
described by:

D = Kch(1.35 r

1.05

[1-128]

9 O i nr

where Krn = an erodibility factor [(lb/ft /s)(ft /lb) " ],T = average shear
(lb/ft2) of the flow in the channel, and Tcr= a critical shear stress (lb/ft2)
below which erosion is negligible. Critical shear stress seems to increase
greatly over the year as the soil consolidates (13) .

Shear stress is assumed to be triangularly distributed in time during the
runoff event in order to estimate the time that shear stress is greater than
the critical shear stress. For the time that shear stress is greater than cri-
tical shear stress, shear stress is assumed constant and equal to peak shear
stress for the storm.

Until the channel reaches the nonerodible layer, an active channel is as-
sumed that is rectangular with the width obtained from figures I - 1 1 and 1-12
and equations [1-130] and [1-131]. The solution requires finding a value of xc.
Given the discharge Q, Manning's n, friction slope Sf, a value g(xc) is calcu-
lated from:

3/8

g (xc)

1.49 Sfl/2

[1-129]

Given a particular value g(xc), a value of xc is obtained from figure 1-11.
Having determined xc, a value for R* = hydraulic radius/wetted perimeter and W*
= width/wetted perimeter is read from figure 1-12. The width of the channel is
then calculated from:

Q n

.49 Sf1/2

3/8

R 5/8

[1-130]

16.00

w° 14.00

12.00

2 10.00

8.00

6.00

4.00

2.00

0.00

GEOMETRIC PROPERTIES OF
ERODED CHANNELS AT EQUILIBRIUM

00 .10 .20 .30 .40 .50

X (DISTANCE ALONG WETTED PERIMETER
FROM WATER SURFACE DOWN TO POINT
WHERE LOCAL SHEAR STRESS EQUALS
CRITICAL SHEAR STRESS) DIVIDED BY
WETTED PERIMETER

Figure I — 11 .—Function g(xc) for equilibrium
eroded channel

The functions shown in figures I — 11 and 1-12 are. stored piecewise in the model.
The channel moves downward at the rate dcn:

1.05

[1-131]

dch = Gm/Psoil = Kch t1'35 T " Tcr) ' 'Psoil
where em = erosion rate calculated using the maximum shear stress (mass/unit
area/unit time), and psoii = mass density of the soil in place. The erosion
rate in the channel is:

J. 05

Ech = Wac Kch(1.35

[1-132]

where Ecn is the soil loss per unit channel length (mass/unit channel length/
unit time).

0.80

0.60-

0.40

0.20

000

. 1 — 1~— ^^ — 1 1 1 1 1 1 1

^^(Wp= WETTED PERIMETER)

^^ N. y~ AREA = AREA/wJ

^\ ^\X HYD. RAO =HYD. RAD/Wn .

^v: WIDTH =WIDTH/Wn\.

r-DEPTH¥= DEPTH/Wp ^ — ~\^

GEOMETRIC PROPERTIES OF ^V. \ .
ERODED CHANNELS AT EQUILIBRIUM \\

-0.12

)8 2

0.04

0.00

0.00 0.10 0.20 0.30 0.40 0.50
Xr (DISTANCE ALONG WETTED PERIMETER FROM WATER SURFACE
c DOWN TO POINT WHERE LOCAL SHEAR STRESS EQUALS

CRITICAL SHEAR STRESS) DIVIDED BY WETTED PERIMETER

Figure 1-12.—- Equilibrium eroded channel geometric properties

Once the channel hits the nonerodible boundary, the erosion rate begins to
decrease with time. The width W of the channel at any time after the channel
has eroded to the nonerodible layer is estimated by:

oo = 1 -exp (-t*)
where oj is the nondimensional channel width given by:

to = (W - Wi)/(Wf - W-j)
t = (t - ti)(dw/dt)i/(Wf - Wi)

[1-133]

[1-134]
[1-135]

where W-j = width at t = t-j, W = width at t, Wf = final eroded width for t ■> «
and the given Q, t = time, and (dw/dt)-,- = rate that channel widens at t = t-j.
The initial widening rate is given by:

1.05

(dwVdt)i = 2 Kch (xb - Tcr)i-UD/pS0il

where Tb ">s 9"iven by:

and

1/2
f(xD) = V* = Tm[(8 xb) -2 xb]

T m = Tmax/^ = ^ -^5

[1-136]

[1-137]
[1-138]

where xb = flow depth/wetted perimeter, and Tmax = maximum shear stress at cen-
ter of channel .

The final width Wf is determined by finding the xcf that gives:
3/8 YS- ,

__AJ1

1 .49Sf 1/2

(1 - 2xcf)3/8f(Xcf)

xcf

[1-139]

where f(xcf) is the function given by equations 1-137 and 1-138 and evaluated
at xcf. The final width is:

f |_1.49Sfl/2j

3/8

l-2x

xCf

5/3

3/8

[1-140]

Sediment Transport and Partitioning of Shear Stress

Sediment transport capacity for the channel is described using the Yalin
equation in exactly the same form as it was used in the overland flow element.

The shear stress acting on the soil is the shear stress used to compute
detachment and transport. Grass and mulch reduce this stress. Total shear is
divided into that acting on the vegetation or mulch and that acting on the soil
using sediment transport theory (13) .

First, velocity is estimated using n^, the total Manning's n [See (3),
(26) for estimates of n^-]. The hydraulic radius due to the soil is:

Rsoil = (V nbch/Sfl/*)

3/2

[1-141]

where nbcn = Manning's n for a bare channel and Sf = friction slope. Shear
stress acting on the soil is:

Tsoil

YRsoil sf

Tcov = Y Sf[V(nt - nbch)/1.49 Sf1/2]372.

[1-142]
[1-143]

If xcov exceeds the shear stress at which the cover starts to move, the cover
fails, thereby increasing the flow's shear stress on the soil.

Variations in parameters such as Manning's n and slope along the channel
can be considered. In addition, the model breaks the channel into segments of
length Leff/10. Calculations begin at the upper end of the channel and proceed
downstream.

IMPOUNDMENT (POND) ELEMENT

The impoundment (pond) element describes deposition behind impoundments
(including parallel tile outlet) that drain following each storm.

Deposition is the main sedimentation process occurring in impoundments.

Since transport capacity in the impoundments is essentially nonexistent, the
amount of sediment trapped in an impoundment is basically a function of time
available for sediment to settle to the bottom of the impoundment before flow
leaves the impoundment. The equations for the pond element were developed from
regression analyses that fit relationships to output from a more complex model
(15, 16) which had been previously validated with field data.

The fraction of particles of a specific size and density that passes
through the impoundment is:

fpi = Ai exp(B! • deqi) [1-144]

where fpi- = fraction passing through pond for particle type i, A]_ and B]_ = co-
efficients given below, and dGq-j = the equivalent sand diameter of particle
type i (microns). The particle types in the model represent classes rather
than specific particles. Therefore, equation [1-144] was integrated over the
class range and divided by the class width to obtain average for the class as:

Fpi = Ai [exp(B!du) - exp (Bid1)]/(B1 • Ad) [1-145]

where Fp-j = fraction passed for particle class i. The equivalent sand diame-
ters are arranged in ascending order, and du is the dGql- for the class and d]_
is the next smallest dgq-j . The diameters d uand d^are not centered around
deqi because dGq -j is assumed to represent the maximum diameter in the class.
The class width Ad = du - dj. Values of Fp-j are limited to a maximum of 1.0.
The coefficients Ai and B^ are given by:

Ax =1.136 exp(Zs) [1-146]

Bi = -0.152 exp(Ys) [1-147]

with Zs and Ys in turn given by:

Zs = (-6.68 x 10~6)f - 0.0903B+ (1.19 x 10"4)Cor [1-148]

-(3.42 x 10"6)Vr - 204001

Ys = (3.28 x 10"5)f + 0.123B -(2.4 x 10"4)Cor [1-149]

+(8.10 x 10~6)Vr -118801

where f and B = coefficient and exponent in a power equation relating surface
area to depth Sa = fyp B , yp = depth in pond (ft), Sa = surface area (ft^), Vr
= volume of runoff (ft^), and I = infiltration rate in the pond (ft/s). The
coefficient Cor is related to the orifice in the pipe outlet by:

Cor = 13968 dor2 [1-150]

where dor = diameter of the orifice (ft). Also, the coefficient Cor is related
to discharge and the depth above the outlet point by:

Cor = 3600 Qp/yd1/2 [1-151]

where Q = discharge (ft /s) and yj = depth (ft).

All of the water which enters the pond will not leave. The volume leaving
is estimated by:

Vout = 0.95 Vin exp(Zr) [1-152]

where Vout = volume of runoff discharged, V-jn = volume of runoff reaching the
pond, and Zris given by:

[1-153]

Zr = - (9.29 x 10~6) f + 0.282B + (1.25 x 10"4)Cor
- (3.08 x 10"6) Vr - 333591

[1-154]

If Vout > Vin, Vout = Vin [1-155]

are additional constraints on Vout for equation [1-152].

VALIDITY OF THE MODEL

Comparison with Other Models

The validity of the model can be partially assessed by comparing it with
other models that might be used in this application. The detachment relation-
ships used in the overland flow element gave good results for a watershed at
Treynor, Iowa. Estimates were considerably better than those from the USLE
using storm EI (12J and better than those obtained from a procedure using run-
off volume and peak discharge alone as an erosivity factor (25). Both rainfall
and runoff appear to be important for estimating detachment on overland flow
areas. More comprehensive models like ARM (6_) or ANSWERS (I) use modifications
of the USLE and/or require data for calibration. The CREAMS erosion/sediment
yield component preserves the USLE form when erosion is simulated for a range
of storms and slope lengths and steepnesses. On long-term simulation, the mod-
el produces results comparable with those of the USLE. Information to select
overland flow erosion parameters is as readily available for CREAMS as it is
for the USLE.

Comparison of Output from Model with Observed Data

The validity of the model has been partially assessed by comparing output
from the model with measured sediment yield from concave field plots under sim-
ulated rainfall, single terrace watersheds, small watersheds with impoundment
terraces, and a small watershed with conservation tillage. The simulations
were made using measured rainfall and runoff values. Parameter values were
selected from volume II, chapter 2 without calibration, except as noted.

Concave Plots

Three concave plots 35 ft long were carefully shaped in a soil where soil
properties were uniform within the depth of shaping. Slope along the plots
continuously decreased from 18% at the upper end to 0% at the lower end. Simu-
lated rainfall at 2.5 in/hr was used to detach and provide runoff to transport
sediment (vol. Ill, ch. 10). The measured particle distribution of the sedi-
ment reaching the deposition area was used as input to the model. The soil
erodibility factor and Manning's n were adjusted in the model to give observed
soil loss entering the deposition area at the lower end of the plots. The
estimated sediment yield for the 29-ft plot was 0.0026 lb/ft/s compared with
0.0017 lb/ft/s observed. For the 35-ft plot, the estimated and observed values
were 0.0014 and 0.00094 lb/ft/s, respectively.

Single Terrace Watersheds

Soil loss was simulated for 8 yr of data from small, single terrace water-
sheds at Guthrie, Okla. (4). The simulations were made without calibration.
Table I -10 gives computed and measured results.

Table 1-10. — Comparison of simulated sediment yield from single terrace water-
sheds with observed values

Tprr,rp firadp Sediment yield

lerrace brade Observed Simulated

(tons/acre) (tons/acre]

2B Variable, 0.0033 at outlet to 0.0 54 28

at upper end.

3B Variable, 0.005 at outlet to 0.0 62 53

at upper end.

3C Constant, 0.005. 54 47

5C Constant, 0.0017. 21 20

Impoundment Terraces

Soil loss was simulated for selected storms representing a range of rain-
fall and runoff characteristics for three locations in Iowa; Eldora, Charles
City, and Guthrie Center, from an impoundment terrace study (17). The model
was run without calibration. The results are given in table I — 11 .

Small Watershed

Simulations were run without calibration for approximately 2-1/2 years of
data from the P2 watershed at Watki nsvi lie, Ga. in a conservation tillage sys-
tem for corn (_27 ) . Deposition in the backwater from the flume at the watershed
outlet was modeled. Deposition measured in the flume backwater was about equal
to the measured sediment yield on a similar nearby watershed (19). The com-
puted total sediment yield for the period of record was 6.6 tons/acre, while
the measured value was 8.3 tons/acre.

Table 1-11. — Summary of observed and simulated sediment yield from impoundment

terraces in Iowa

Watershed

Area

Julian
date

Sediment

yield

Observed

Simulated

(acres)

(lb)

Charles City

4.6

70147

1,197

70152

70244

160

70323

71151

280

294

71157

209

160

Eldora

1.8

68198

283

150

68220

69187

1,057

554

69232

124

227

71163

335

139

Guthrie Center

1.4

69207

256

273

69249

70144

122

70162

198

123

70167

70229

Overland Flow Sediment Transport

Estimates for sediment transport capacity of overland flow may be in error
by a factor of two (vol. Ill, Ch. 10). However, the sediment transport equa-
tions used by other models have not been tested against field data where depo-
sition was known with certainty to be limiting sediment load. Overland flow
conditions are outside the range of most sediment transport equations developed
for streamflow, and consequently, many give results greatly in error for over-
land flow (vol. Ill, ch. 10). Given the present state of the art, the trans-
port relationship used in this model is believed to be as adequate as any
available, especially when the equation is not calibrated.

Channel Erosion

The channel erosion relationships are the ones most likely to be in error
even though they fit data from a rill erosion study wery well (vol. Ill, ch.
11). Data from the rills (12 in wide) may not scale up to channel size (that
is, 10 ft wide). However, computed final channel width agreed well with
observed widths for a wide range of streams. While the channel erosion rate
for a single storm may be in error, the upper limit for annual channel erosion
should be a reasonable for soils having a nonerodible layer beneath the soil
surface.

Proven parameter values for the channel soil erodibility and critical

shear stress are not available. CREAMS considers the decay in erosion with

time due to previous erosion; most models do not, with the exception of Bruce's

and others {2) . This component of CREAMS require calibration.

Backwater

Most erosion models as applied to fields use a kinematic runoff simulation
model to generate values for hydraulic variables. That is, friction slope is
set equal to the channel slope. This prevents modeling deposition in a back-
water area at the field outlet. Such deposition occurs often and is important
in estimating chemical yields associated with enrichment of fine sediment dur-
ing deposition. The solutions to the spatially varied flow equations discussed
earlier account for these outlet controls, and thus can be used to simulate
sediment deposition.

SIMULATION COSTS

Comprehensive models that simulate erosion over space and over time
through a runoff event are potentially more powerful than this model. However,
detailed downslope spatial variability (slope, cover, and so forth) can be
analyzed with this model. The expected slight increase in improved estimates
with a more comprehensive model probably does not offset the additional costs
for computing, and moreover, many of these models require lumped parameter
estimates which prevents their consideration of slope shape and buffer strips,
for example, that can be analyzed with this model.

While computer costs vary from site to site and change often, rough esti-
mates are, nontheless, important for qualitative comparisons. Using the CDC
6500 Computer at Purdue University, simulation costs for the erosion/sediment
yield component of CREAMS were about $0.10 per storm event. Therefore, the
erosion/sediment yield component can simulate individual storm events for a
cost of about $1 to $3 per year. Although the model is quite comprehensive,
the programming is efficient and simulation costs are not prohibitive.

SUMMARY

An erosion/sediment yield model for field-sized areas was developed for
use on a storm-by-storm basis. The overall objective was to develop a model,
incorporating fundamental erosion/sediment transport relationships, to evaluate
best management practices. Although the procedure does not consider changes in
parameter values within individual storms, it does allow these parameters to
change from storm to storm throughout the season. Moreover, parameters of the
model allow for distribution of field characteristics along overland flow
slopes and along waterways. Many of the model parameters are selected using
tested methods developed for the well-known Universal Soil Loss Equation. For
this reason, we feel that the model has immediate applications without the need
for extensive calibration.

Limited testing has shown that the procedures developed herein give
improved estimates over the USLE and modified USLE procedures. Specific com-
ponents of the model were tested using experimental data from overland flow,
erodible channel, and impoundment studies. Sensitivity analyses are described
in chapter 6. Application of the model is demonstrated in volume II, chapter
2. Initial results suggest that the model produces reasonable results and is a
useful tool for analyzing the influence of alternate management practices.

REFERENCES

(1) Beasley, D. B. , E. J. Monke, and L. F. Huggins.

1977. The ANSWERS model: A planning tool for watershed research.
American Society of Agricultural Engineers, Paper No. 77-2532. St.
Joseph, Mich.

(2) Bruce, R. R. , L. A. Harper, R. A. Leonard, W. M. Snyder, and others.

1975. A model for runoff of pesticides from small upland watersheds.
Journal of Environmental Qua! ity 4(4) : 541 -548.

(3) Chow, V. T
)59. (
York,~WT. 680 pp.

1959. Open-Channel Hydraulics. McGraw-Hill Book Company, Inc., New

(4) Daniel, H. A., H. M. Elwell, and M. B. Cox.

1943. Investigation in erosion control and reclamation of eroded land
at the Red Plains Conservation Experiment Station, Guthrie, Oklahoma,
1930-40. U.S. Department of Agriculture, Technical Bulletin No.
837.

(5) Davis, S. S.

1978. Deposition of nonuniform sediment by overland flow on concave
slopes. M.S. Thesis, Purdue University, West Lafayette, Ind.

(6) Donigian, A. S. , Jr., and N. H. Crawford.

1976. Modeling nonpoint source pollution from the land surface. U.S.
Environmental Protection Agency, EPA-600/376-083. 279 pp.

(7) Einstein, H. A.

1968. Deposition of suspended particles in a gravel bed. Journal of
the Hydraulics Division, Proceedings of the American Society of Civil
Engineers 94(HY5) : 1197-1205.

(8) Foster, G. R.

1979. Sediment yield from farm fields: The Universal Soil Loss Equa-
tion and onfarm 208 plan implementation. J_n Universal Soil Loss
Equation: Past, Present, and Future, Chapter 3. Soil Science Socie-
ty of America. Madison, Wise.

(9) , and L. F. Huggins.

1977. Deposition of sediment by overland flow on concave slopes. J_n
Soil Erosion Prediction and Control. Soil Conservation Society of
America Special Publication No. 21, pp. 167-182. Ankeny, Iowa.

(10) Foster, G. R., and L. D. Meyer.

1972. Transport of soil particles by shallow flow. Transactions of
the American Society of Agricultural Engineers 15(1) :99-102 .

(11) , and L. D. Meyer.

1975. Mathematical simulation of upland erosion by fundamental erosion
mechanics. J_n Present and Prospective Technology for Predicting
Sediment Yields and Sources. U.S. Department of Agriculture, Agri-
cultural Research Service, Southern Region, ARS-S-40, pp. 190-207.
(Service discontinued; Agricultural Research Service is now Science
and Education Administration-Agricultural Research.)

(12) , L. D. Meyer, and C. A. Onstad.

1977. A runoff erosivity factor and variable slope length exponents
for soil loss estimates. Transactions of the American Society of
Agricultural Engineers 20(4) :683-687.

(13) Graf, W. H.

1971. Hydraulics of Sediment Transport. McGraw-Hill Book Co., New
York, NY. 544 pp.

(14) Khaleel, R., G. R. Foster, K. R. Reddy, M. R. Overcash, and others.

1980. A nonpoi nt source model for land areas receiving animal wastes:
III. A conceptual model for sediment and manure transport. Transac-
tions of the American Society of Agricultural Engineers. (In press.)

(15) Laflen, J. M., and H. P. Johnson.

1976. Soil and water loss from impoundment terrace systems. In Pro-
ceedings of the Third Federal Inter-Agency Sedimentation Conference,
Water Resources Council, Chapter 2, pp. 303-41, Washington, D.C.

(16) , H. P. Johnson, and R. 0. Hartwig.

1978. Sedimentation modeling of impoundment terraces. Transactions
of the Amererican Society of Agricultural Engineers, 21 (6) : 1131-1135 .

(17) , H. P. Johnson, and R. C. Reeve.

1972. Soil loss from tile outlet terraces. Journal of Soil and Water
Conservation 27(2) :74-77 .

(18) Lane, L. J., D. A. Woolhiser, and V. Yevjevich.

1975. Influence of simplification in watershed geometry in simulation
of surface runoff. Colorado State University, Hydrology Paper No.
81, Fort Collins, Colorado. 50 pp.

(19) Langdale, G. W., A. P. Barnett, R. A. Leonard, and W. G. Fleming.

1979. Reduction of soil erosion by no-till systems in the Southern
Piedmont. Transactions of the American Society of Agricultural
Engineers 22(l):82-86, 92.

(20) Li, R. M.

1977. Water and Sediment routing from watersheds. Proceedings of
River Mechanics Institute, Colorado State University, Fort Collins,

Colorado. Chapter 9.

(21) Lombardi, F.

1979. Universal Soil Loss Equation (USLE), runoff erosivity factor,
slope length exponent, and slope steepness exponent for individual
storms. PhD Thesis, Purdue University, W. Lafayette, Ind.

(22) Mantz, P. A.

1977. Incipient transport of fine grains and flakes of fluids - ex-
tended Shields diagram. Journal of Hydraulics Division, Proceedings
of the American Society of Civil Engineers 103(HY6) :601-615.

(23) McCool, D. K., W. R. Gwinn, W. 0. Ree, and J. E. Garton.

1966. Spatially varied steady flow in a vegetated channel. Transac-
tions of the American Society of Agricultural Engineers 9(3) :440-444.

(24) Onstad, C. A., and G. R. Foster.

1975. Erosion modeling on a watershed. Transactions of the American
Society of Agricultural Engineers 18(2) :288-292.

(25) , R. F. Piest, and K. E. Saxton.

1976. Watershed erosion model validation for Southwest Iowa. I_n Pro-
ceedings of the Third Federal Inter-Agency Sedimentation Conference,
Water Resources Council, Chapter 1, pp. 22-24. Washington, D.C.

(26) Ree, W. 0., and F. R. Crow.

1977. Friction factors for vegetated waterways of small slopes. U.S.
Department of Agriculture, Agricultural Research Service, Southern
Region, ARS-S-151, 56 pp. (Series discontinued; Agricultural Re-
search Service is now Science and Education Administration-Agricul-
tural Research.)

(27) Smith, C. N., R. A. Leonard, G. W. Langdale, and G. W. Bailey.

1978. Transport of agricultural chemicals from upland Piedmont water-
sheds. U.S. Environmental Protection Agency, EPA-600/3-78-056 .

(28) Smith, R. E.

1977. Field test of a distributed watershed erosion/sedimentation mod-
el. In Soil Erosion: Prediction and Control. Soil Conservation
Society of America, Special Publication No. 21, pp. 201-209, Ankeny,
Iowa.

(29) Williams, J. R.

1975. Sediment -yield prediction with universal equation using runoff
energy factor. J_n Present and Prospective Technology for Predicting
Sediment Yields and Sources, U.S. Department of Agriculture, Agricul-
tural Research Service, Southern Region, ARS-S-40, pp. 244-252.
(Series discontinued; Agricultural Research Service is now Science
and Education Administration-Agricultural Research.)

(30) , and H. D. Berndt.

1977. Determining the Universal Soil Loss Equation's length-slope fac-
tor for watersheds. J_n Soil Erosion: Prediction and Control. Soil
Conservation Society of America, Special Publication No. 21, pp.
217-225, Ankeny, Iowa.

(31) Wischmeier, W. H. , and D. D. Smith.

1978. Predicting rainfall erosion losses. U.S. Department of Agricul-
ture, Agriculture Handbook No. 537, 58 pp.

(32) Yalin, Y. S.

1963. An expression for bedload transportation. Journal of the Hy-
draulics Division, Proceedings of the American Society of Civil Engi-
neers 89(HY3):221-250.

(33) Young, R. A.

1978. Review of eroded sediment particle size and density data. Per-
sonal Correspondence. U.S. Department of Agriculture, Science and
Education Administration, Morris, Minn.

Chapter 4. THE NUTRIENT SUBMODEL
M. H. Frere, J. D. Ross, and L. J. Lane^'

INTRODUCTION

Nutrients are naturally occurring chemicals essential for plant growth.
A total of 16 chemical elements are necessary for the growth and reproduction
of most plants, although the most significant are nitrogen (N), phosphorus (P),
and potassium (K). Most soils are deficient in N, P, and K for optimum plant
production, and thus commercially-available fertilizers contain these nutrients
essential to maintain the current level of agricultural production. The other
nutrient elements may be added as impurities in the fertilizer or applied to
treat specific nutritional problems. Present evidence indicates that nitrogen
and phosphorus are the principal nutrient pollutants and, therefore, only these
nutrients are considered in this model.

A major source of nutrients reaching water bodies in this country is sew-
age, both from municipal treatment plants and nonsewered residences. These
represent point sources of pollution and extensive efforts are underway to
limit their contributions. Runoff from rural land is another major source, but
unlike point sources, runoff integrates the contribution from a diffuse, dynam-
ic source. It must be recognized that some nutrients leave the system even
when fertilizer is not applied and while we cannot eliminate all nutrient
losses, it is desirable to minimize them.

It must be emphasized at the beginning that the dynamic system under con-
sideration is complex. The wide variety of climates and landscapes provides
such a wide range of results that there is no typical case. Complications are
introduced by difficulties in chemical analysis for nutrients in water samples.
Numerous procedures have been followed for chemical analysis, but changes in
nutrient form can occur between the times of sampling and sample analysis.
Some nutrient data have been reported as the soluble form when actually they
could have been associated with colloidal material not removed from the sample.
The practical significance of these complications is unknown, but they are
noted at this time to alert the reader of the limitation of the data associated
with nutrient pollution.

THE PROBLEMS

Two problems are associated with nutrients in the aquatic environment:
(1) Water may be toxic to humans, animals, or fish when the concentration of

1/ Soil scientist, USDA-SEA-AR, Southern Region Office, New Orleans, La.;
mathematician, USDA-SEA-AR, Durant, Okla.; and hydrologist, USDA-SEA-AR, Tuc-
son, Ariz., respectively.

certain nutrient forms exceeds a critical level, and (2) eutrophication may be
accelerated. The nitrite form of nitrogen, which is the most toxic, interacts
with components in the blood to interfere with oxygen transport. Methemoglo-
binema, the technical name of this illness, is often called "the blue baby syn-
drome" because infants are very susceptible. Most of the problems with drink-
ing water have been associated with farm wells with faulty well casings, loca-
ted close to manure concentrations such as barnyards.

Nitrate is 5 to 10 times less toxic than nitrite. Children convert some
nitrate to nitrite in their stomachs and can develop methemoglobinemia. The
U. S. Drinking Water Standards set the limit for nitrate at 10 mg nitrogen/1
and recommendations for livestock are 10 times higher. Dissolved ammonia is
another form of nitrogen that can occur at levels toxic to fish. Micro-organ-
isms can generate free ammonia from organic matter in lake bottoms during sum-
mer stagnation periods. Trout are sensitive to 1 to 2 ppm ammonia while gold-
fish appear to be less sensitive.

Eutrophication is the enrichment of waters by nutrients and the ensuing
luxuriant growth of plants. Rapid growth of algae is the greatest and most
widespread eutrophication problem in most states. Algae can create obnoxious
conditions in ponded waters, increase water treatment costs by clogging screens
and requiring more chemicals, and cause serious taste and odor problems. When
a large mass of algae dies and begins to decay, the oxygen dissolved in the wa-
ter decreases and certain toxins are produced, both of which may kill fish.

Aquatic plants require a number of nutrients for growth, but nitrogen and
phosphorus appear to be the ones accounting for most of the excessive growth.
Eutrophication appears to become a problem when the concentration of inorganic
nitrogen exceeds about 0.3 ppm and inorganic phosphorus exceeds about 0.015
ppm. These concentrations of inorganic forms of nutrients are maintained by
microbial conversion of organic forms so the total input of nitrogen and phos-
phorus per unit area of the lake (loading rate) is important. Current interna-
tional quidelines for eutrophication control are 2.0 to 5.0 kg of P and 50 to
100 kg of nitrogen per surface hectare of lake per year (1.8 - 4.5 lb P/acre
and 45 - 90 lb N/acre) .

CONTROL OF NUTRIENT POLLUTION

Nutrients as related to water quality are transported from the watershed
by three processes: runoff, erosion, and leaching. Soluble forms of nitrogen
and phosphorus are transported in the runoff. Insoluble forms and forms adsor-
bed to sediment particles are moved by erosion. Nitrate is the principal nu-
trient form leached to groundwater or base flow by percolating water.

To reduce the concentration of nutrients in receiving water (amount of nu-
trient per unit volume of water) or the total amount (load), either the amount
of nutrient available for transport or the transport process must be reduced.
The amount of nutrient available for transport can be reduced by practices such
as applying fertilizers, manures, and wastes when the runoff, erosion, and
leaching processes are at a minimum, or by incorporating the nutrients into the
soil so that they are not accessible to runoff water. Conservation practices
such as contour farming, conservation tillage, terracing, and grassed waterways
can reduce the amount of runoff or erosion or leaching.

Unfortunately, each practice has its limitations and a combination of
practices are often needed. In addition, a practice that controls one problem
may induce another. As an example, practices reducing runoff may create leach-
ing problems. Finally, we must recognize that the effect of practices cannot
be evaluated by a single storm event. Few, if any, storms produce the same re-
sults. Consequently, a practice or combination of practices must be evaluated
for different types and sizes of storms occurring at different stages of growth
or times of the year.

NUTRIENT MODEL INPUT AND OUTPUT

Weather, soils, topography, and land use all effect the performance of a
conservation or pollution control practice, and a comprehensive mathematical
model is useful in the evaluation. As shown in figure 1-13, the model requires
information about hydrology, erosion, and particular nutrient characteristics
of the field to predict the nitrogen and phosphorus moving in runoff, with ero-
sion, and by leaching. The hydrology model provides estimates of the volume of
runoff, percolation, soil water and temperature, and plant growth and water-
use, while the erosion model provides estimates of sediment loss on a field
scale. The model predicts the average concentration of soluble N and P in the
runoff. Multiplying the average concentration by the volume of runoff esti-
mates the total amount or load produced by the storm. The model provides an
estimate of the amount of nitrate leached and its average concentration. The
estimates of N and P associated with sediment from erosion corresponds to the
total N and total P often reported in water quality studies.

HYDROLOGY
MODEL

N8P

RUNOFF

NUTRIENT
DATA

NUTRIENT
MODEL

NUTRIENT
LEACHING

N8P

WITH

SEDIMENT

Figure 1-13. — Flow diagram of input and output
for the nutrient model.

SEDIMENT TRANSPORT OF NUTRIENTS

The loss of total N and total P from cropland ranges from about 1 to 50 or
100 kg/ha per year. Sediment can be a major transport vehicle for phosphorus
and organic nitrogen. Raindrop splash and flowing water detach soil particles
and organic matter containing nitrogen and phosphorus. The transport capacity
of the flowing water depends primarily on the volume and velocity of water
flow. Whenever the velocity is reduced, such as by a flatter slope, the trans-
port capacity is reduced and any sediment in excess of the reduced capacity
settles out. Since larger and heavier particles settle out first, the remain-
ing sediment contains a larger percentage of finer particles which have a high-
er capacity per unit of sediment to absorb phosphate and organic nitrogen.
Also organic matter is lighter and tends to be associated with the fine parti-
cles. Thus, the transported sediment is richer in phosphorus and nitrogen than
the original soil. Figure I-14(a) is a flow diagram for this process in the
nutrient model .

The sediment of concern in this model is limited to that from surface
rill and interrill erosion. Sediment from gully, channel, or other sources of
erosion is not included in the calculation because the nutrient content of the
soil changes significantly with soil depth. In fact, some subsoils high in
clay content may absorb phosphate and can deplete the solution concentration of
phosphate from the soil.

(a)

EROSION

ENRICHMENT

Nap

WITH
SEDIMENT

(b)

ENRICHMENT

SEDIMENT

Figure 1-14. — (a) Diagram for estimating nitrogen and
phosphorus losses with sediment; (b) relation of
enrichment to amount of sediment.

Algorithm

The kilograms/hectare of nitrogen or phosphorus transported by sediment
(SEDN or SEDP) is predicted in this model by the following equation:

and

SED- = SOIL- * SED * ER- [1-156]

ER- = A- * SED ** B- [1-157]

where SOIL- is the N (SOILN) or P (SOILP) content (kg/kg soil) in the field,
SED is the kg/ha of sediment predicted by the erosion model. ER- is the en-
richment ratio for N or P, A- is a coefficient for N or P, and B- is an expo-
nent for N or P.

Soils typically contain 0.05 to 0.3 percent nitrogen and 0.01 to 0.13 per-
cent phosphorus (see vol. Ill, ch. 13, table 1). This sixfold to tenfold range
indicates that a measurement of the specific field involved is highly desir-
able. Applications of fertilizers, manures, wastes, and crop residues increase
the N and P content above natural levels while intensive cropping without nu-
trient additions reduces the N and P content (vol. Ill, ch. 15). Many samples
of soils are analyzed each year by State and commercial soil testing laborator-
ies. Adequate data for a specific field might already be available from this
source or could be measured in a short period of time. If no other information
is available, approximate values (vol. Ill, ch. 14 and 15) could be used, rec-
ognizing the error that can be made.

The enrichment of sediment, as described above, occurs because of selec-
tive erosion and deposition processes. An evaluation of all available data in-
dicates that the logarithmic relation, figure I-14(b), between the enrichment
ratio and the amount of sediment holds for wide ranges of soil and vegetative
conditions (vol. Ill, ch. 12). It is suspected that changing soils, crops, or
management practices should result in different coefficients and exponents in
the equation. However, the amount of data available at the present time does
not permit a statistically significant distinction in these parameters. Con-
siderable research work is presently being conducted that should be useful in
improving this relation. Using a value of 7.4 for A- and -0.2 for B-, the en-
richment ratio for both nitrogen and phosphorus can be predicted within a fac-
tor of two for an annual average and a factor of five for individual storm
events. The implications of these relations for sediment transport of nutri-
ents is that reducing sediment transport will not reduce nutrient transport by
an equal amount. Conservation practices that reduce erosion will reduce nutri-
ent transport but to a smaller degree.

SOLUBLE NUTRIENTS IN RUNOFF WATERS

Runoff waters contain soluble forms of nitrogen and phosphorus ranging
from 0.01 to 1 ppm P and 0.1 to 10 ppm N with loads from less than 0.1 to as
much as 10 kg/ha/yr or lb/acre/yr (figures 1-15 and 1-16). This is one of the
most difficult areas to model because of the variety of nutrient sources and
processes of extraction. While considerable data have been reported on the

CONCENTRATION ppm
O.OI OJ LO IO0 1 00.0

]P PRECIPITATION (ZZZZZZJN

CROPLAND RUNOFF Dl EZZZZZZZZZZZZ] N

Y////////////A N NON-CROPLAND RUNOFF

D DRAINAGE K///////IN

Figure 1-15. — Range of nitrogen and phosphorus concentrations in
different waters.

SPATIAL RATE Ibs/acre/yeor OR kg/ha/year
0,01 OJ I.Q 10.0 100.0

PRECIPITATION PI I EZZZZZZZZZN

^ZZZ2nc

ROPLAND RUNOFF

1 | ^"non-cropland runoff

PI I DRAINAGE CZZZZZZZZN

CROPLAND SEDIMENT p\>///////////// /////7777\N

PI 1

[/////////////IN

NON-CROPLAND SEDIMENT

Figure 1-16. — Range of spatial rates of nitrogen and phosphorus in
waters and sediments.

integrated or gross effects, very little research has been reported on the in-
dividual processes (vol. Ill, ch. 14 and 15). Figure 1-17 illustrates the con-
cepts used in this model for predicting the soluble forms of N and P in runoff
waters.

N IN
RAINFALL

RESIDUES
FERTILIZERS
SOLID WASTE

SOLUBLE

NSP IN Icm

SOIL

RUNOFF

INFILTRATION

N LEACHED

DEEPER INTO

SOIL

Nap

IN
RUNOFF

Figure 1-17. — Diagram for estimating nutrient losses in runoff,

Rainfall

Rainfall is the driving force for the system and it also contains nutri-
ents. The concentrations of nutrients in rainfall not only vary across the
country (figure 1-18) but within short distances and during a storm. Most of
the phosphate is associated with dust and is generally neglected as an input.
Nitrate and ammonium are the principal nitrogen forms occurring in precipita-
tion and their sum averages from 1 kg of N/hectare/yr in the West to over 3 kg
of N/hectare/yr in the Great Lakes area (0.9 - 2.7 lb N/acre/yr) . This level
is not agronomically significant for cropland but could be for unfertilized
range and forest areas. Seasonal charts indicate the highest concentrations
occur in the spring and summer. Nitrogen concentrations are often found to be
slightly higher during the first part of a storm.

The concentration of nitrogen in rain ranges from a little less than 1
ppm to a little over 1 ppm. This concentration range corresponds to the lower
end of the concentration range for runoff from cropland and the upper end for
runoff from noncropland (figure 1-15).

Applied Nutrients

Other sources of nutrients are the fertilizers, manure, and plant residues
placed on the surface. With the exception of slow-release fertilizers,

Figure 1-18. — Nitrogen contributions (NO3-N and NH3-N), kilograms per hectare
per yeart from rainfall throughout the United states [from (1)].

nitrogen fertilizers are quite water soluble and phosphate fertilizers are mo-
derately soluble. Consequently, water from the soil and light rains dissolves
the granules. Most fertilizer and manure applied to cropland is mixed into the
soil by plowing or disking leaving only a proportionate fraction near the sur-
face.

Plant materials are not usually "applied" to a field, rather they are the
living crop canopy or the residues left after the harvest of the crop. They
include the stems and leaves of such crops as small grains, corn, sorghum, soy-
beans, and cotton. On rangeland, pasture, hay meadows, and cropland with sod
crops, there is a significant amount of litter and mulch on the soil surface
besides the living crops. This plant material contains organic forms of nitro-
gen and phosphorus, some of which can be leached from the residue during
storms. Plant material changes the nitrogen content of runoff on the order of
+ 10 percent but increases the phosphorus content as much as fourfold (vol.
Ill, ch. 15). Table 1-12 gives some estimates for amounts of residues and
their nutrient content. These values can vary by a factor of two across the
country and the fraction of the nutrient content that can be leached out is
probably 50% + 20%. Concentrations on the order of 0.1 ppm P have been ob-
served in washoff from mature cotton plants.

Table 1-12. Approximate yield and nutrient content of selected crops
can vary by a factor of two across the country)

values

Crop

rield

Nitrogen

Phosphorus-

Alfalfa^

(kg/ha)
8,960

(units/acre)

(kq/ha) (1
224

3/acre)

(kg/ha) (1
20

Vacre)

(4 tons)

(200)

(18)

Barley

grain

2,150

(40 bu)

( 35)

' 6)

2/
Beans^7

straw

2,240

(1 ton)

( 15)

( 2)

(dry)

1,950

(30 bu)

( 75)

(10)

Bermudagrass

17,920

(8 tons)

224

(200)

(30)

Bluegrass

4,480

(2 tons)

( 60)

( 8)

Cabbage

44,800

(20 tons)

168

(150)

(16)

Clover %!

red

4,480

(2 tons)

( 80)

(10)

white

4,480

(2 tons)

146

(130)

(10)

Corn

grain

9,400

(150 bu)

151

(135)

(24)

stover

10,080

(4.5 tons)

112

(100)

(16)

silage

56,000

(25 tons)

224

(200)

(30)

Cotton 1 int & seed

2,240

(1 ton)

60)

(12)

2/
Cowpea hay-

stalks

2,240

(1 ton)

( 45)

( 6)

4,480

(2 tons)

134

(120)

(10)

Lettuce 2/
Lespedeza-

44,800

(20 tons)

100

( 90)

(12)

4,480

(2 tons)

( 85)

( 8)

Oats

grain

3,200

(90 bu)

55)

(10)

straw

4,480

(2 tons)

( 25)

[ 8)

Onions

16,800

(7.5 tons)

45)

Oranges^ ,
Peanut sA'

62,720

(28 tons)

( 85)

12)

nuts

3,360

(1.5 tons)

123

(110)

Potatoes

tubers

44,800

(400 cwt)

106

95)

12)

vines

2,240

(1 ton)

100

( 90)

Rice

grain

4,540

(90 bu)

55)

12)

straw

5,600

(2.5 tons)

; 30)

Rye

grain

1,880

(30 bu)

35)

4 (

straw

3,360

(1.5 tons)

: 15)

Sorghum

grain

3,360

(60 bu)

50)

10)

2/
Soybean-

stover

6,720

(3 tons)

65)

grain

3,020

(45 bu)

179

160)

16)

straw

2,240

(1 ton)

25)

4 (

Sugarbeets

roots

44,800

(20 tons)

85)

14)

tops

26,880

(12 tons)

123

110)

10)

Sugarcane

stalks

67,200

(30 tons)

112

100)

20)

tops

29,120

(13 tons)

56 (

50)

11 (

10)

Timothy

5,600

(2.5 tons)

60)

11 (

10)

Tobacco

3,360

(1.5 tons)

129

115)

11 (

10)

Tomatoes

fruit

56,000

(25 tons)

162

145)

22 (

20)

vines

3,360

(1.5 tons)

78 (

70)

11 (

10)

Wheat

grain

3,360

(50 bu)

73 (

65)

16 (

14)

straw

3,360

(1.5 tons)

20)

2 (

1/ Pounds P = 0.436 lbs P2O5.

2/ Legumes that do not reguire fertilizer nitrogen.

Animal manures may be applied to crop, pasture, and even rangelands if the
manure is available. The manure is available and even a disposal problem when
animal production is part of the total agricultural unit. Beef production po-
ses different problems than dairy, swine, and poultry operations. Most beef
animals spend part of their lives in an open grazing situation and the rest of
the time in a confined feedlot where the manure can be collected and used. The
nutrient content of manure varies with the animal and the type of feed used.
An average content is given in table 1-13. About 50% of the N is lost during
handling, storage, and application. In addition, only about 50% of the applied
organic N is mineralized and available to plants in the first cropping season.

Table 1-13. — Nutrient content of

manuresl/ The P content remains relatively

stable during handling, storage, and
application. The fraction of the nu-
trient content in manure that can be
leached out is probably about the
same as for plant residues 50% + 20%.

Animals

(%)

Beef

2.5

0.8

Dairy

2.0

Swine

2.8

1.0

Laying hens

4.3

1.3

Broilers

3.8

1.3

Soluble Nitrogen and Phosphorus in
the Surface Soil

The surface layer of soil con-
tains a certain amount of soluble N
1/ On a dry weight basis after and P. Estimating this value along
losses during handling, storage, and with a runoff extracting or efficien-
application. cy factor is the weakest part of the

nutrient model. Fortunately, the
concentration and load of soluble nutrients in runoff are not usually the domi-
nant factor. However, conservation practices that reduce sediment transport of
nutrients, like notill, can increase their concentration in runoff water and
the relative importance of this pathway.

At the present time with our limited understanding of the system, we as-
sume that a 1 cm layer of soil interacts with the rain. All the soluble N and
P are expected to exist in the water of the pores. Only a fraction of the nu-
trients are extracted into the flowing water. Analysis of pesticide and nutri-
ent data suggest that this extraction coefficient ranges from 0.01 to 0.4. The
loss of freshly applied fertilizer is also seldom more than 1 to 5% (vol. Ill,
ch. 15). Based on the range of extraction coefficients and the observed range
of P concentration in runoff, the concentration of soluble P in the surface
layer would range from an upper level of 5 ppm to less than 2 ppm.

The porosity of soils is usually in the order of 40% + 10% and a hectare
would contain 4 x 10^ kg of water in the upper centimeter layer. At 5 ppm this
is equivalent to 0.20 kg/ha. For the lack of better information, we assume
similar concentrations for nitrogen.

Leaching from the Surface Layer

Only part of the rainfall leaves the field as runoff. Frequently there
will be no runoff from a storm. The part of the rainfall that does not runoff

fills the surface layer and leaches soluble nutrients deeper into the soil. In
this model the amount leached is proportional to the fraction of the rainfall
that does not run off. It is subtracted before runoff to account for non-run-
off producing rains. Soluble nitrogen compounds leached into the soil are as-
sumed to be nitrate or quickly converted to nitrate and are added to the ni-
trate pool in the soil. Soluble phosphate compounds are leached out of the
surface layer but do not move on through the soil because of its large buffer-
ing capacity. The buffering capacity of the soil also prevents the phosphate
concentration from dropping below a level characteristic of the soil.

Algorithm

The basic assumption is that the rate of change in concentration of solu-
ble nitrogen in the water in the surface (1 cm) of soil is proportional to the
difference between the existing concentration and the concentration of nitrogen
in the rainfall. That is, we assume

{j| = Kx f(t)(Cr-C) [1-158]

where Ki is a rate constant for downward movement, f(t) is infiltration rate,
Cr is concentration in the rainfall, and C is concentration in the soil
surface. The mean concentration during infiltration is

h = ((c0-Cr)/KlF)(1-exp(-KlF)) + Cr [1-159]

where C0 is initial concentration and F is total infiltration for the storm.
The concentration at the end of infiltration and the start of runoff is

Cl = (VCr)exp(~KlF) + Cr " [1-160]

The final concentration after runoff is

C2 = (CrCr)exp(-K2Q) + Cr [1-161]

and the mean concentration during runoff is

C2 = ((CrCr)/K2Q)(l-exp(-K2Q)) + Cp [1-162]

where K2 is a rate constant for movement into runoff and Q is total runoff.

The extraction coefficients are

EXKN: = d POR Kx [1-163]

for downward movement and

EXKMg = d POR K2 [1-164]

for movement in runoff where d = 10 mm is the depth of the surface layer and
POR is the porosity.

The equations for soluble phosphorus are similar except the rainfall con-
centration Cr is assumed zero but it is replaced by a base concentration Cg
due to the buffering action described earlier.

Downward movement of nitrogen is calculated as

DWN = C1*EXKN1*FI*0.01 [1-165]

where FI is the total infiltration minus an initial abstraction assumed equal
to the volume of pore space in the surface soil (d POR). Similarly, the amount
of soluble nitrogen in runoff is

RON = C2*EXKN2*Q*0.01 . [1-166]

The amount of souble phosphorus is calculated as

ROP = C*EXKP2*Q*0.01 [1-167]

where EXKP2 is the phosphorus extraction coefficient for movement into runoff.

The concentration in the soil solution is 10 times (depth of active sur-
face layer = 10 mm) the kilograms/hectare of soluble N or Ps SOL-, divided by
the porosity, POR. The amount of soluble N or P is increased by the amount in
rainfall and additions of fertilizers, manure, and plant residues, F-, on sev-
eral application dates, DF. These additions are assumed instantaneous inputs
or impulse additions to the soil solution while the rainfall nitrogen is dis-
tributed throughout the storm.

The application factor, FA, is 1 for surface-applied nutrients and equal
to the fraction of the application remaining in the 1 cm surface layer if the
nutrients are incorporated. The initial value of SOL- reflects the soil contri-
bution only. All fertilizers, wastes, and residues are input via additions.
If the addition occurred before the first simulation date, the date of applica-
tion can be set as zero.

In the simplified model described above, soluble nutrients are moved out
of the surface layer with infiltration. Soil evaporation may contribute to the
nitrogen concentration in the surface layer if nitrogen is transported upward
with the water flux due to evaporation. This contribution may be estimated as
the product of the NO3 concentration in the root zone and the water flux due to
evaporation.

Exact values of the nitrogen and phosphorus extraction coefficients are
unknown. For the downward movement coefficients, values of 1.0 imply complete
efficiency for infiltrating water while values equal to the extraction coeffi-
cients for movement in surface runoff imply the same efficiency for downward
movement as for the movement in runoff. As a first approximation, and in the
absence of experimental data, we assumed the downward movement extraction co-
efficients were less than 1.0 but greater than the extraction coefficients for
movement in surface runoff. Simulation runs were made with the downward
movement extraction coefficients varying from 0.1 to 1.0. Values near 1.0 re-
sult in yery rapid decreases in concentration due to infiltration. Values in
the range 0.2 to 0.5 resulted in somewhat higher soil concentrations and more
reasonable concentrations in surface runoff. Therefore, we arbitrarily fixed
the downward movement coefficients at 0.25. Recognizing the interaction be-
tween the extraction coefficients, this assumption forces all the variability
into the extraction coefficients for movement in runoff. However, since the
extraction coefficient is a number reflecting the complex movement of soluble
nutrients, and as such is a simplification, we fixed two of the coefficients
and let the two remaining coefficients represent extraction efficiency. The
constraint is that the extraction coefficients for movement in runoff must be
less than the downward movement extraction coefficients.

NITROGEN CYCLING AND NITRATE LEACHING

One of the important pathways of nitrogen loss is the leaching of nitrate
from the root zone to ground water, tile drains, or base flow. Figure 1-15
illustrates that the concentration of nitrate in drainage can exceed 10 ppm.
In order to predict this source of pollution it is necessary to maintain a bud-
get of nitrate and water in the root zone. The water budget and movement are
calculated in the hydrology model. Figure 1-19 shows the inputs and withdraw-
als for the nitrogen system. The nitrogen cycle in the soil is extremely

SOLUBLE N

FERTILIZER

RESIDUES

WASTES

infiltration

ncorporotion

SOIL

ORGANIC

NITROGEN

soil water ond
temperature

mineralization

TR ROOT
A_ ZONE

^ E

soil water

plant growth

PLANT
NITROGEN

denitrif ication

-« —

}

, percolation

NITROGEN
GASES

NITRATE
LEACHED

Figure 1-19. — Diagram for estimating nitrate leaching

complex and very complex models have been developed to describe the system.
Unfortunately, complex models require numerous parameters which are not usually
available. Therefore, we have chosen to include the minimum number of rela-
tions and parameters that will provide an acceptable estimate of the system be-
havior.

The active root zone of a crop depends upon the crop, the soil, and the
stage of growth. The root zone is important because it determines the depth of
soil from which the plant is removing nitrate. In this model we neglect the
fact that early in the season the root zone is shallower than late in the sea-
son. We use only the final root zone depth because the nitrate is not lost to
potential uptake until it has moved below this zone. The density of roots and
extraction of water and nutrients are concentrated in the upper part of the
root zone. Therefore, while a few roots may grow very deep, the effective root
zone is much shallower. Table 1-14 gives depths of active root zones for some
crops. Soil conditions, like hard pans, sand or gravel layers, or acid subsoil
can often limit the depth of the root zones. Therefore, it is important to
check the characteristics of the field under study.

Table 1-14. — Root zones of various
crops

Crop

Corn

Sorghum

Alfalfa

Tomatoes

Wheat

Sugarbeets
Soybeans
Field beans
Potatoes

Active
root zone

(In)
48

1,200

Soluble N from the Surface Layer

As described in a previous sec-
tion, soluble nitrogen is leached
from the surface layer by infiltra-
ted water into the root zone. All
of the soluble N is nitrate or forms
that are quickly converted to ni-
trate.

Applied Nutrients

900

Fertilizers, wastes, and plant
residues are often injected, plowed

under, disked in, or otherwise in-

Pasture 24 600 corporated into the soil. Not all

Onions 18 450 of the nitrogen in these materials

Bluegrass lawns is in the nitrate form or forms rap-
idly converted to nitrate. However,
for simplicity we have assumed that only nitrate or forms easily converted to
nitrate are included in the application amount. Other forms of N would add to
the organic nitrogen pool which is already very large. In this version of the
model the potentially mi neral izable nitrogen in the soil is not increased by
the applied nutrients.

Soil Organic Nitrogen

Micro-organisms in the soil convert organic forms of nitrogen to nitrate.
This process, mineralization, is sensitive to temperature and moisture condi-
tions in the soil. Only a part of the organic nitrogen is readily converted
and a chemical test of the sample of soil from the field in question is the
best method for determining the potentially mi neral izable nitrogen. In the

event this is not possible, a less exact estimate can be made from the total N
and organic carbon contents for various soil orders (vol. Ill, ch. 13, table
1). Increasing temperature increases the rate of mineralization in an exponen-
tial manner up to a peak at 35°C (308° kelvin, 95°F). The optimum moisture
content for mineralization is "field capacity" (when grapitational water has
drained away, about 1/3 bar tension). Mineralization decreases linearly with a
decrease in water content below this optimum.

Algorithm

The following equations are used to calculate the kilograms/hectare of
mineralized nitrogen, MN, during a period, DAYS, between storm events:

TK = EXP(15.807 - 6350/TA) [1-168]

WK = AWC/FC [1-169]

MN = P0TM*WK*(1-EXP(-TK*DAYS) [1-170]

where TK, the temperature coefficient, is calculated from the average tempera-
ture during the period, TA, in degrees kelvin; WK, the water coefficient, is
calculated from the average volumetric water content during the period, AWC,
and the field capacity, FC, both in fractions (mrTp/mm^); and POTM is the kilo-
grams/hectare of potentially mineral izable nitrogen in the soil. The average
temperature and the average soil water content are supplied by the hydrology
model. While the temperature of the soil in the root zone is preferred, air
temperature can be used because the error introduced is small compared with
other sources of error.

The potential mineral izable nitrogen, POTM, should reflect agricultural
practices such as the amount of plant residue. However, for a single cultural
practice, POTM can be initialized each year (for example, at the date of crop
emergence) to the original measured value for the soil.

Plant Nitrogen

Crop growth is the most important process that removes nitrate from the
root zone. Under good weather conditions and with good management practices,
most of the nitrate in the root zone is taken up by plants. In this model two
options are provided for calculating plant uptake of nitrogen.

Algorithms

Option I simulates plant growth as a function of plant water use and N up-
take as a function of plant N content. Accumulated dry matter production
(grain + stover + roots) is calculated with the equation

DMi = £!5ii (YP)(K) [1-171]

PWU

where DM is accumulated dry matter production on day i, WU is daily plant water
use, PWU is total plant water use for the growing season, YP is the crop yield
potential, and K is the ratio of dry matter to crop yield at maturity.

The N poncentration in plants is a function of plant maturity as expressed
by the equation

d= MINIMUM MDMi/TDM)^ D_172]

b3(DMi/TDM)

where c-j is the N concentration in the plant on day i, TDM is the total dry
matter production for the growing season, and b\ , b2 , b3, and b4 are parameters
defined by Smith and others, table 3 (vol. Ill, ch. 13).

The accumulated N uptake can be computed for any day during the growing
season with the equation

UNn- = c-j DMi [1-173]

where UN is the accumulated N uptake for day i. The N uptake during any period
of the growing season is simply the difference between beginning and ending
accumulated amounts.

Option II assumes nitrogen uptake follows a normal probability (S shaped)
curve which is reduced for moisture stress. The equations are:

PUN = 1 - 1/2 (S)-4 [1-174]

S = 1.0 + 0.196854X + 0.115194X2 + 0.000344X3 + 0.01957X4 [1-175]

X = (T-M)/SD [1-176]

where PUN is the fraction of the potential annual plant nitrogen taken up by T
days of growth. M is the days of growth required to take up 50% of the annual
amount. SD is the number of days between 50% and 84% uptake and corresponds to
1 standard deviation. When X is negative, PUN = 1-PUN, the lower part of a
symetrical curve. The amount of nitrogen uptake between storms, UN, kilograms/
hectare, is:

UN = (PUN-PPUN)*PU*TR [1-177]

where PPUN is the previous level of uptake at the last storm; PU is the poten-
tial annual nitrogen uptake for the crop in kilograms/hectare; and TR is the
ratio of actual to potential transpiration during the period.

This option should be most useful if concentration parameters used in
equation [1-172] are not available.

Nitrate Leaching

The amount of nitrate leached below the root zone is the fraction of the
water in the root zone that percolates out of the root zone times the amount of
nitrate in the root zone. The amount of percolation between storms is calcu-
lated by the hydrology model. The amount of water remaining in the root zone
after percolation is the field capacity times the depth of the root zone.

Most nitrate leaching occurs in the winter and spring when plants are not
extracting much water. October 1 is usually the date when the soil profile is
driest and that date (Julian day 274) is used in this model as the initial day
for accumulating the annual amounts of water percolated and nitrate leached.
Thus, DRAIN is the accumulated amount of percolated water from each storm event
and TOTNL is the accumulated amount of nitrate leached.

Algorithms

One estimate of the leaching fraction, Method A, assuming complete mixing,
is:

FL = PERC/(PERC + RZC) [1-178]

where PERC is the mm of percolation and RZC is the mm of water remaining in the
root zone. The kilograms/hectare of nitrate leached after each storm, NL, is
FL * N03 where N03 is the kilograms/hectare of nitrate in the root zone at the
time of the storm.

There is a second way, (Method B in vol. Ill, ch. 13), to estimate the
annual amount of nitrate leached. This method calculates the leaching frac-
tion, FL, as

FL = (DRAIN/(DRAIN + 10FC))**EX [1-179]

where EX = (RZMAX - 300)/10 [1-180]

where DRAIN is the annual depth of percolation in millimeters, RZMAX is the
root zone depth in millimeters, and FC is the water fraction of the soil at
field capacity. The constant, 300, assumes that the nitrate was uniformally
distributed in the top 600 mm (2 ft) of the soil.

Denitrification

Under anaerobic conditions in the soil, nitrate can be reduced to nitrogen
gases. The process is considered to be a first-order reaction sensitive to or-
ganic carbon, temperature, and moisture. In this simple model, the rate con-
stant at 35°C is calculated from the amount of organic matter in the soil and
adjusted for temperature assuming a twofold reduction for each ten degrees de-
crease in temperature. The effect of moisture is accounted for by permitting
denitrification only for the number of days of drainage in excess of a half
day; that is, when the moisture content of the soil exceeds field capacity.

Algorithm

The amount of soil carbon is calculated from the amount of organic matter:

SC = 0M/0.1724 [1-181]

where SC is milligrams carbon/gram of soil and 0M is percent organic matter.
The rate constant, DK, at 35°C is calculated according to:

DK = 24*(0.011 * SC + 0.0025) [1-182]

where DK has units of day" . The temperature adjusted rate constant is:

DKT = exp (0.0693 * ATP + DB) [1-183]

with

DB = In DK - 2.4255 [1-184]

and ATP is equal to the average temperature, degrees Celsius, used in the min-
eralization process. The amount of denitrification, DNI, in kilograms/hectare,
between storms is:

DNI = N03 * (l.-exp(-DKT*(DT-.5)) [1-185]

where N03 is the kilograms/hectare of nitrate in the root zone and DT is the
number of days of drainage since the last storm. The half day subtraction is
justified on the basis that many short drainage periods have only short periods
of anaerobic conditions.

TESTING THE NUTRIENT SUBMODEL

A limited test of the submodel was made using data from watershed P2 at
Watkinsville, Ga. Nitrogen losses in runoff, with sediment, and in percolation
below the root zone and phosphorus losses in runoff and with sediment were sim-
ulated for 1974. The simulated losses in runoff and with sediment were compar-
ed to observations (3.) • Fertilizer (N-P-K) was applied at the rates of 38-33-
127 kg/ha and incorporated to an average depth of 10 cm. Corn was planted
immediately after fertilization. Forty-three days after planting 101 kg/ha of
N was applied to the surface soil by spray.

Model input values for the simulation are given in table 1-15. The poten-
tial N uptake for Option 2 was estimated from the sampled grain and stover
yield. Values for the parameters in table 1-15 were taken from Smith and
others (3) .

The climatic and hydrologic data given in table 1-16 are, except for aver-
age soil water content, measured values. The transpiration ratio was set at

75%.

Measured and simulated runoff and sediment losses of nitrogen and phos-
phorus are compared in table 1-17 and by relating observed and computed yields
for individual storms. The relation between computed and observed soluble ni-
trogen in runoff was

N = -0.04 + 0.89 N [1-186]

R2 = 0.94

where Nq is computed nitrogen yield and Nq is observed nitrogen yield. For
soluble phosphorus the regression equation was

Table 1-15. — Model input values used for simulation, watershed P2 at Watkins-

ville, Ga.f 1974

Input
parameters

Parameter description

Parameter
values

Initial conditions

SOLN

SOLP

N03

SOILN

SOILP

Nutrient parameters

EXKN

EXKP

POTM

RCN

Nutrient additions

Soluble nitrogen in the surface cm (kg/ha)
Soluble phosphorus in the surface cm (kg/ha)
Nitrate in the root zone (kg/ha)
Soil nitrogen associated with sediment (kg/kg)
Soil phosphorus associated with sediment (kg/kg)

Extraction coefficient of nitrogen into runoff
Extraction coefficient of phosphorus into runoff
Enrichment coefficient of nitrogen in sediment
Enrichment exponent of nitrogen in sediment
Enrichment coefficient of phosphorus in sediment
Enrichment exponent of phosphorus in sediment
Potentially mineral izable nitrogen (kg/ha)
Nitrogen concentration in rain (ppm)

.00035

.00018

.075

.16

.146

NF
DF

Number of additions
Date of appl ication

Amount of nitrogen in application (kg/ha)

Amount of phosphorus in application (kg/ha)

Fraction of application in top cm

Nitrogen uptake

OPT

Common to hydrology

DEMERG

DHRVST

RZMAX

P0R

Plant growth simulation
Potential yield of crop (kg/ha)
Coefficients for the nitrogen concentration
in the plant.

Date of crop emergence

Date of crop harvest

Depth of potential root zone (mm)

Porosity (mm^/mm^)

Water content at field capacity (mm/mm)

Organic matter (%)

74119

74174

102

.1
1.0

5700

.0209

- .157
.0128

- .415

74125

74303

450

,45
= 20
,65

Table 1-16. — Climatic and hydrology data for Watershed P2, Watki nsvil le, Ga.

Average

temperature

Average

Rain-

Run-

Perco-

between

soil water

Date

fall

off

lation

events

content

Sediment

Model param-
eter

DATE

RAIN

RUNOFF

PERC

ATP

AWC

SED

Month,

Year and

day

Jul i an day

(mm)

(°c)

(mm/mm)

(kg/ha)

April 4

74094

33.0

3.0

29.11

0.200

9.6

April 12

74102

1.0

14.83

.195

April 13

74103

24.0

4.0

13.13

15.83

.192

14.5

April 22

74112

8.0

3.26

16.67

.197

May 2

74122

2.0

18.17

.195

May 4

74124

9.0

1.55

19.28

.192

May 5

74125

19.0

1.0

17.56

19.56

.200

10.1

May 11

74131

3.0

20.11

.198

May 12

74132

13.0

12.02

20.67

.200

May 15

74135

3.0

.66

21.00

.199

May 23

74143

70.0

7.0

58.25

21.89

.197

92.0

May 26

74146

7.0

4.54

22.94

.199

May 31

74151

13.0

9.89

23.61

.199

June 8

74159

8.0

3.84

24.39

.197

June 10

74161

6.0

4.42

24.89

.200

June 20

74171

12.0

1.0

4.21

25.39

.196

1.4

June 27

74178

108.0

42.9

57.82

25.83

.196

966.5

July 17

74198

3.0

26.33

.189

July 23

74204

3.0

26.56

.182

July 24

74205

15.0

1.0

2.08

26.56

.184

23.4

July 26

74207

13.0

12.32

26.56

.200

July 27

74208

72.0

45.6

26.30

26.56

.200

661.3

Aug 5

74217

1.0

26.44

.193

Aug 7

74219

27.0

18.06

26.38

.189

Aug 10

74222

28.0

2.0

25.03

26.33

.200

22.6

Aug 14

74226

8.0

5.95

26.28

.199

Aug 16

74228

51.0

8.0

39.47

26.17

.199

70.7

Aug 17

74229

15.0

1.0

13.51

26.06

.200

7.3

Aug 29

74241

17.0

1.0

2.58

25.67

.189

3.8

Sept 1

74244

11.0

1.0

7.82

25.06

.199

Sept 3

74246

8.0

6.86

24.89

.199

Sept 6

74249

23.0

21.38

24.56

.199

Sept 25

74268

4.0

22.78

.193

Oct 16

74289

9.0

18.61

.188

CT>co«3<r)iHOr--moiMo

HOlOHfOMNSHOO
OOOf-OLDOOOOO

rHOHl^CO^MlDOlflVD

<£>t£>c\jr-.ir>voc\jou->.— ir-H

0*OWr- I «3" I— 1*3-000

Ot-io<Ttonooooo

Lf><T>ir>«3-<noooovoror-»cr>

MCOOOl>!CHNS«tHH
ONOtOOlfiOWOOO
OOO^HO^HOOOOO

CVJr-H^J-r-lCriCOCVJ'd-VOLnVO

OOOOOOOOOOOOr-H
OOOCOOOOOOOO

CO CM UJCOlDHOlrHNO

*± CM I MlO<tOWOiOM

*i- <d I O00 00 00HWWO

O CM ICT^OCOOCMOOO

HtO CM .-H CM t-H LT> LT> CO CO

o> 00 icvj.— icsir-io^i-csjcsj

O f-~ I CTiCONiDlOMrHH

00 it-~ooo>-tooo

O .-I

«* n en <j- in o o iv «> cm <j
omocON^oOrHr-icftco
.-ic\Jcoc\j.-icn«3-r-)C\j»-(C\j

OHHinMNOHOOO

O r-< ,-1

iocsiti-^ot-i^t-focricocric\j

cotr>cocovDooio«=rvoo 1— i

OHuniflCOMlD^OCVJHH CM

OCMOCOOOOr-HOOO LT>

O t-H .-<

rocMCMCsJCMOvor-^cn

ID CM

0)
>>>>£=

CM 1—

£Z r— ^— C7> C7> C7> C7> CL
33333333CU

,~3ro,-orDcc<:"=c«a;oo

PQ = 0.02 + 0.40 PQ [1-187]

R2 = 0.48

where Pq is computed phosphorus yield in runoff and Pq is the corresponding ob-
served value.

Nutrients transported with sediment were predicted for individual storms
and related to observed values. Predicted and observed nitrogen with sediment
were related as

Ns = 0.05 + 0.81 N [1-188]

R2 = 0.92

where Ns is predicted and Ns is observed nitrogen yield with sediment. The
corresponding equation for phosphorus was

P = 0.02 + 0.82 Ps [1-189]

R2 ■ °-91

where Ps is predicted and Ps is observed phosphorus yield with sediment.

With the exception of soluble phosphorus (equation [1-187]) the model pre-
dicted total yields and reproduced trends in the observed data. In this case
the regression coefficient of 0.40 suggests a bias in the predicted phosphorus
yields with runoff. However, the extraction coefficients for movement of ni-
trogen and phosphorus in runoff were obtained by calibration with the observed
data while the remainder of the parameters were estimated prior to the simula-
tion runs.

The computed nitrogen balance for 1974 on watershed P2 at Watkinsville,
Ga. is summarized in table 1-18. Soil N refers to the soluble nitrogen content
in the surface 1 cm of soil and soil NO3 refers to the NO3 content in the root
zone. The nitrogen balance is maintained since the change in storage equals
the difference between the input and output. The data shown in table 1-18 do
not include nitrogen losses with soil loss. As shown in table 1-17, these
losses amount to over 4 kg/ha.

SUMMARY

A simplified nutrient model has been developed for field-sized areas. The
simplified model predicts nutrient losses in runoff, in leached water, and with
soil loss. These predictions are not as accurate as desired but they may be
within the limitations of our capability at this time. Additional testing and
evaluation are described in chapter 6.

Table 1-18. — Nitrogen balance for watershed P2 near Watkinsvil le, Ga., 1974

Source
or
process

Input

Output

Change in storage

(kg/ha)

(kg/ha

Rainfall

,18

Fertilizer

140

Mineralization

43,

.13

Soil N

Soil N03

Runoff

3.76

Leaching

33.98

Dentrification

37.83

Uptake

124.66

(kg/ha!

+ 0.30
- 9.23

Total

191.3

200.2

8.9

REFERENCES

(1) Chapin, J. D., and P. D. Uttormark.

1973. Atmospheric contribution of nitrogen and phosphorus. The Univer-
sity of Wisconsin Water Resources Center, Hydraulic and Sanitary Labo-
ratory, Technical Report WIS-WRC 73-2. Madison, Wis.

(2) Chichester, F. W. and S. J. Smith.

1978. Disposition of 15N-Labeled fertilizer nitrate applied during corn
culture in field lysimeters. Journal of Environmental Quality 7(2):
227-233.

(3) Smith, C. N., G. W. Bailey, R. A. Leonard, and G. W. Langdale.

1978. Transport of agricultural chemicals from small upland Piedmont
watersheds. U.S. Environmental Protection Agency, EPA-600/3-78-056 .

Chapter 5. THE PESTICIDE SUBMODEL
R. A. Leonard and R. D. Wauchope-

The pesticide submodel was developed on simplified concepts of processes
and designed to be responsive to different management options. Foliar- and
soil- applied pesticides are separately described so that different decay rates
can be used for each source of the same chemical if necessary. Usually
pesticide residing on foliage dissipates more rapidly than that from soil.
Also decay rates can be made site-specific if information is available.
Movement of pesticides from the soil surface as a result of infiltrating water
is estimated using differences of rainfall and runoff for the storm and
pesticide mobility parameters. Pesticide in runoff is partitioned between the
solution or water phase and the sediment phase. This aspect is particularly
important when examining management options that limit sediment yield.
Comprehensive discussions are provided in volume III, Supporting Documentation,
to aid the user in assigning appropriate parameter values.

THE RUNOFF SYSTEM

A simple conceptualization of the runoff system is shown in figure 1-20.
The primary source of pesticide available to enter the runoff stream is visual-
ized as a surface layer of soil defined arbitrarily as having a depth of 1 cm.
This definition is based on observations by Leonard and others (_10) that runoff
concentrations of both dissolved and adsorbed pesticides were strongly
correlated with pesticide concentrations in this layer. Actually the thickness
of this layer depends on many factors. Pesticide extraction by raindrop splash
and interrill soil movement may occur in a very shallow layer, whereas
extraction from rills may extend several centimeters deep. In models by Bruce
and others (2J and Frere and others (_5) , rill and interrill extractions were
described separately, but here the process was conceptually combined for
simplicity. Others have defined this effective thickness to be about 0.5 cm
(3) and 2.5 cm (12).

Washoff of pesticide applied to foliage is another source that may enter
the runoff stream. In this model, the fraction of applied pesticide intercep-
ted by foliage is specified initially. Dislodgeable residue remaining on the
foliage at the time of rainfall is estimated from information given in volume
III, chapter 18. The fraction of this dislodgeable residue removed by rainfall
is then added to the soil surface 0 to 1 cm zone and a new concentration for
this zone is computed for the runoff event.

1/ Soil scientist, USDA-SEA-AR, Southeast Watershed Research Program,
Athens, Ga., and chemist, USDA-SEA-AR, U.S. Delta States Agricultural Research
Center, Stoneville, Miss., respectively.

SOURCE OF
PESTICIDE
FOR RUNOFF
(SOIL SURFACE
ZONE)

DEPTH OF SOIL
INCORPORATED
PESTICIDE

Figure 1-20. — Schematic representation of the
conceptualized runoff process.

Pesticide dissipates from the surface zone primarily by degradation and
volatilization processes. During rainfall events, pesticide may move below the
surface zone in the infiltrating water and across the surface in runoff. In
the model, initial concentrations of unincorporated pesticides are computed as
if they were uniformly incorporated into the 0 to 1 cm depth. Concentrations
of incorporated pesticides are computed based on their incorporation depth and
efficiency of incorporation. A simplified schematic of the pesticide submodel
is shown in figure 1-21.

DESCRIPTION OF THE PESTICIDE SOURCE

As stated previously, the source zone for extraction into runoff was arbi-
trarily defined as the 0 to 1 cm depth increment of the soil surface. Concen-
trations are computed in units of micrograms/gram or parts per million. For
pesticides applied directly to the soil surface the concentration resulting
from the application, C, , is:

R x

[1-190]

PESTICIDE

APPLICATION

(R OR CQ)

FRACTION ON SOIL
(C,)

FRACTION ON FOLIAGE
(M,)

ADD INITIAL RESIDUES

(c2)

ADD INITIAL RESIDUES
(M2)

COMPUTE

CONCENTRATION

OF RESIDUE

(c4)

COMPUTE

MASS

OF RESIDUE

(M4)

ADJUST FOR
DOWNWARD MOVEMENT

WASHOFF
FRACTION

RAINFALL,
RUNOFF, SEDIMENT
(HYDROLOGY AND
EROSION MODELS)

(FROM OTHER MODELS)

COMPUTE
AVAILABLE

CONCENTRATIONS
IN WATER

RES

FOR <

DUE

5TORM

AND SEDIMENT AND
TOTAL MASS

Figure 1-21. — Simplified schematic representation of the pesticide model

where BD is the bulk density of the surface soil layer, and R is the applica-
tion rate in units of kilograms/hectares. Assuming an average BD of 1.5,

Cx = R x 6.7. [1-191]

For soil incorporated pesticides,

Cx = 6.7 R x EF/ID [1-192]

where EF is a unitless factor to compensate for nonuniform incorporation and ID
is the incorporation depth. If uniform mixing is assumed, EF = 1; however, ex-
perience has shown that uniform mixing is rarely achieved (11) . Concentrations
in the surface 0 to 1 cm layer are usually higher than computed assuming uni-
formity so that EF probably ranges from 1 to 3. In situations where pesticide
is injected or banded below the soil surface, EF may be less than 1. A range
of 0.5 to 1 is suggested. Normally EF would be assigned a value of 1 unless
information is available on the incorporation pattern in a specific situation
of interest. If some pesticide residue, C2 , was initially present in the soil
at the time of application, the total or net concentration would be C3 = C2 +
Cl.

When pesticides are applied to foliage, the areal concentration expressed
in units of milligrams/square meters is

Mj = R x FF x 100 [1-193]

where FF is the fraction of the application intercepted by the foliage. M is
not concentration on the leaf surface, but a concentration based on the pro-
jected ground area. Unless the canopy is dense with complete closure, a frac-
tion of the application, SF, will also be intercepted by the soil surface.
Soil concentration resulting from this application is computed as above as C\ -
R x SF. When aerial applications are made, losses by drift and volatilization
may occur so that FF + SF will not equal 1. Information on foliar interception
is found in volume III, chapter 18.

Residues of the same pesticide from previous applications, if present in
either the soil or foliage compartment, are added to that resulting from the
new application for the total residue level. At the beginning of the model ap-
plication period, any initial residues present are specified. When pesticide
residues are redistributed in the soil by major tillage, a new model applica-
tion period should be begun, with the resultant surface concentration input as
initial residue for this period. The surface concentrations at the beginning
of the period may be estimated from the residue remaining and the tillage
depth.

DISSIPATION OF PESTICIDE FROM THE RUNOFF-ACTIVE ZONE

A simple exponential dissipation rate is assumed for both soil and foliar
residues throughout the model application period. For soil residue, C4, the
concentration remaining at time, t, in days after application of the pesticide
or in days after specifying the concentration of initial residue is:

C4 = C3e"kst. [1-194]

Likewise, mass remaining on foliage, M., at time, t, is:

M4 = M3e"kft, [1-195]

/0.693 t\

M4 = M3e^ Cl/2 i [1-196]

where Z\j2 is the "half-life" or half-concentration time of the foliar residue
in days. In the model, the mass of foliar pesticide of concern is that "dis-
lodgeable" or potentially removed by rainfall. Rate constants, ks, for dissi-
pation from soil are tabulated in volume III, chapter 17 for selected
pesticides based on reviews of published data, along with discussions on their
use and limitations. Half-lives of foliar residues are given in volume III,
chapter 18.

FOLIAR WASHOFF -- CONTRIBUTIONS TO RUNOFF

Little information is available in patterns of pesticide removal by rain-
fall. In the model, the assumption is made that once rainfall exceeds a thresh-
old value corresponding to the amount that can be retained as droplets on the
canopy, the fraction potentially dislodgeable is removed during the event (see
vol. Ill, ch. 18, for estimates of percent dislodgeable for selected pesti-
cides). This amount is then added to the soil pesticide residue present at the
time of the event. For computation of concentrations consistent with the con-
ceptual thickness of the soil surface, this mass is distributed evenly in the 0
to 1 cm zone. In reality, washoff may occur during the storm such that foliar
contributions may fall directly into the runoff stream and be transported off
the field. Also, spatial patterns of washoff are likely not uniform and wash-
off may fall into rills under the plant formed by previous rainfall. There-
fore, the assumptions made may tend to underestimate the foliar contribution to
runoff.

Some pesticides, particularly dust formulations, may reach the soil sur-
face by dry-fall between runoff events. Also, drip losses from heavy dew may
remove pesticide from foliage.

VERTICAL MOVEMENT OF PESTICIDE DURING RAINFALL

Runoff potential of mobile pesticides is reduced as infiltrating water
moves some of the pesticide below the soil surface (_1, 4). Pesticide mobility
in soil has been studied extensively using thin layer chromatography technique
(8_, 9_) . With this technique, mobility is expressed relative to the movement of
water (Rf values). In volume III, chapter 19 of this publication, Rf values
are related to Kj, a coefficient describing distribution of pesticide between
the solution phase and the soil phase, defined as a constant for a simple lin-
ear adsorpiton isotherm as:

Kd - -p- [1-197]

where at equilibrium Cs is the concentration, micrograms/gram, in the soil or
solid phase and Cw is the concentration in solution, micrograms/milliliter.
Other procedures for estimating Kj for a number of common pesticides in soil,
along with limitations and possible inherent errors in its use, are also dis-
cussed in chapter 19, volume III.

The following algorithm was developed to estimate vertical movement of
pesticide from the soil surface.

The rate of change of pesticide mass, Z, in the soil surface is

- dZ = C • f • dt [1-198]

where Cw is the pesticide concentration in water or mobile phase and f is the
water flux. At saturation,

Z = Cw- p + Cs (1 - p) D [1-199]

where p is the soil porosity, Cs is the concentration of the adsorbed or immo-
bile phase, and D is the particle density. Introducing Cs = K(jCw and rearrang-
ing the equation above becomes:

c« ■ p+Dyi-p)' "-200]

The rate equation can now be written:

- dZ ■ p + DKd(l-p) dt t1"20"

and integrated between limits of Z0, t0 and Z, t to yield

DKd(l-p) + p)

Z = Zoe

[1-202]

where Z0 is the mass of pesticide present per unit volume of soil surface at
the beginning of the storm. The water flux through the surface during a storm
is

f = RF " *° : S [1-203]

where RF is the amount of rainfall, R0 is runoff, S is the surface storage or
initial abstraction to reach saturation, and t is the storm duration. Making
the substitution for f, t can be eliminated so that

RF - RO - S

DKd(l-p) * P, [,.204]

The value of S is estimated from porosity and the average soil water content

plus canopy stored water. In the model, C4 x BD = Z0 and Z = C/\y x BD where

C/\y is the runoff available pesticide concentration and C4 is as previously de-
scribed.

Where pesticide is foliar applied, the amount assumed to reach the soil by
washoff is added to the surface pesticide residue before estimation of vertical
translocation. This method provides only a crude approximation of the process
compared to other methods (6_, 9J . However, it is developed for use where only
total storm rainfall and runoff amounts are available. Its primary function is
to reduce surface concentrations of those compounds with high soil mobility.
Since the amount of vertical translocation will be small in a single storm for
relatively insoluble compounds, this calculation is by passed in the model if
the compound solubility is < 1 yg/g.

PESTICIDE EXTRACTION INTO RUNOFF

The source zone supplying pesticide to runoff was defined as the surface
(0 to 1 cm) depth increment. At the time of runoff, this increment of soil
contains a pesticide residue specified in the model as the concentration of
"available residue." This is the concentration computed using the appropriate
decay functions, adding any foliar washoff, and allowing for vertical trans-
location. The concentration units are expressed in micrograms/gram of dry soil
as is the convention when a soil sample is removed and analyzed for its pesti-
cide content.

Pesticide is extracted by water flowing over the surface and by dispersion
and mixing of the soil material by the flow and by raindrop impact. Instantan-
eous pressure gradients at the surface caused by raindrop impact on a water-sa-
turated soil could also contribute to exchange of pesticide between the soil
water and the flowing water. At the interface between the soil matrix and the
runoff stream, some mass of soil is "extracted" or is effective in supplying
pesticide to some volume of runoff. The mass of pesticide in this mass of soil
is:

Y = B ' CAV [1-205]

where B is the soil mass per unit volume and C/\v is the concentration of avail-
able residue. As this soil mass mixes or "equilibrates" with the runoff stream

Y = (Cw ' V) + (Cs ' B) [1-206]

where Cw is the pesticide concentration in the water, V is the volume of water
per unit volume of runoff interface, and Cs is the pesticide concentration re-
maining in the soil or solid phase. Ignoring the volume occupied by the soil
mass compared to the larger volume of water; that is, V = total unit volume of
runoff interface = 1,

1 • Cw + Cs • B = B • CAV. [1-207]

By assuming that the distribution between the solution and the soil is approxi-
mated by the equilibrium expression:

K H = -^ , [1-208]

C... - , ± al , [1-209]

'w 1 + B Kd

In these expressions it can be seen that when K<j = 0, then Cw = B C/\y; e.g., if
100 g of soil containing 1 yg/g of pesticide that partitions completely to the
solution phase is extracted by or is dispersed in 1 liter of water, Cw =
100 pg/1 . Also, as K^ becomes large, Cs = C/\y. The numerical value of the
parameter, B, in the above equation cannot be obtained by direct measurement,
and probably is dependent on runoff conditions. However, it will be shown
later that the value ranges from 0.05 to 0.2, with 0.1 giving adequate fit in
most situations.

As material flows from the field, it is assumed that the pesticide concen-
tration in the runoff solution is equal to Cw. However, not all the affected
soil material will become sediment at the field edge. The coarser soil materi-
al will be deposited or left in place. As a result, the transported soil will
have a higher per unit mass adsorptive capacity and adsorbed pesticide concen-
tration than that of the whole soil. Therefore, an enrichment factor is re-
quired and is provided by computations in the erosion submodel (see
vol. I, ch. 3).

Total storm loads are computed as: mass in solution phase = C • storm
:f
yield.

runoff volume, and mass in sediment phase = Cs • enrichment factor • sediment

The approach taken by these procedures differs from other models in that
that the runoff stream is not forced to equilibrate at the soil /water ratio de-
termined by the composition of the saturated soil matrix (5_) nor at a ratio de-
termined by the concentration of the transported sediment, assuming sediment
has the same adsorptive capacity as the soil ( 3J . The weakest assumption,
probably, is that associated with using Kj to partition between the solution
and the soil phase. In addition to the limitations discussed in volume III,
chapter 19, the runoff proccess is dynamic, and true equilibration is probably
never reached. Also, pesticide apparently partitions differently depending on
time of contact with the soil (11); that is, the "apparent Kd" based on observ-
ed partitioning in runoff from experimental watersheds differs from the
laboratory determined values and increases throughout the observation period.

For this reason, Kj may be best used to differentiate between behavior of pest-
icide classes, with K<j ranges differing perhaps by orders of magnitude, that
is, 1 to 10, 10 to 100, and so forth.

TESTING AND EVALUATION

The submodel was tested using data from several experiments conducted un-
der widely different conditions and on pesticides with different properties.
Observed rainfall, runoff, and sediment yield were used as available. No at-
tempt was made to calibrate the submodel or adjust parameters for best fits.
Parameters were estimated either from site-specific information reported or
from information provided in volume III, chapter 16 through chapter 19. As-
signments based on subjective judgment or experience are indicated with expla-
nations. Usually where only a limited number of storms were examined after a
single pesticide application, the computations were made with a desk calcula-
tor.

Table 1-19 lists the parameters estimated for a simulated runoff experi-
ment with lindane and dieldrin under simulated rainfall conducted at Watkins-
ville, Ga., on a Cecil sandy loam soil of about 6% slope (A. W. White, 1970,
unpublished). A total of 6.35 cm of rainfall was applied in 1 hr by procedures
described in White and others (_15 ) . Rainfall was applied on 1, 8, and 28 days
after pesticide application. Runoff was about 50% of applied rainfall. An en-
richment ratio of 1.5 was used based on observations of sand, silt, and clay in
the eroded soil material. Experience has shown that an extraction ratio of 0.2
is required for extreme conditions as prevalent in this experiment. The pesti-
cides decay constant were computed from analysis of soil samples taken during
the experiment. The K(j values were estimated from information in volume III,
chapter 19.

Results of the simulation are given in table 1-20. In general, predicted
compared to observed were reasonably close. Although these compounds ^re not
used to a significant extent presently, these data illustrate the difference in
behavior in runoff as a result of differences in pesticide adsorption (Kd),
solubility, and mode of application which were adequately described.

White and others {15) conducted a similar exeriment with atrazine at 1 hr
and 96 hr after pesticide application. Samples were taken throughout the run-
off event so that discharge-weighted mean concentrations can be computed for
different portions of the runoff event. These data are useful in illustrating
how the model responds to different sized storms. In the model, no upper limit
for single storm loss is used. Since atrazine is somewhat mobile (see vol.
Ill, ch. 19), however, storms of increasing size reduced the computed concen-
trations of runoff-available pesticide, and thereby reduced the predicted run-
off concentrations (table 1-21). The model reduces the concentration of run-
off-available pesticide by an amount proportional to that lost in runoff only
at the end of each event. Therefore, for very large storms where runoff losses
become a significant vehicle for surface depletion, the model may overpredict
total losses, as apparently happened in the prediction for the 6.35-cm storm
(table 1-21).

Table 1-19. — Inputs and parameters for lindane and dieldrin simulation, Cecil
sandy loam soil, Watkinsvil le, Ga.

Lindane

Die

ildrin

Input

Surface

Incorporated

Surface

Incorporated

Application rate, kg/ha.

11.

11.4

Incorporation depth, cm.

7.5

Incorporation efficiency.

Fraction on foliage.

Fraction on soil .

Initial foliar residue,

mg/m^.

Initial soil residue, yg/g

Foliar washoff threshold, cm. 0

Washoff fraction.

Water solubility, ppm.

.12

Foliar residue half-life,
days.

Enrichment ratio.

1.5

Extraction ratio.

Decay constant, ks.

.046

.015

,02

.01

Distribution coefficient,
Kd.

500

Atrazine simulations were compared with observations by Hall (7) on atra-
zine runoff from small plots in Pennsylvania. Values for the decay constant,
ks, and K<j were estimated from information in volume III, chapters 17 and
19. An enrichment factor of 2 was arbitrarily chosen. Simulations gave rea-
sonable predictions; however, the model underestimated solution concentration
in the first event (table 1-22). Predicted concentrations in sediment were
close to observed except for the last two events. This problem apparently is
associated with using Kj as a constant throughout the season. Other data dis-
cussed later show an even greater discrepancy associated with partitioning
pesticide between water and sediment.

In table 1-23 comparisons are made of predicted y_s observed concentrations
of 2,4-D from a treated rangeland watershed (L. J. Lane, 1978, unpublished
data). Acceptable predictions were obtained throughout the observation period.

Table 1-20. — Lindane and dieldrin in runoff; comparison of observed-' vs. predicted values

Pesticide

Time after
application

Concentration in
0-1 cm soil

Concentration in
runoff

Observed Predicted Observed Predicted

Concentration in

sediment

Observed Predicted

(days)

(ug/g) (wg/g)

(mg/l) (mg/1)

(ug/g)

(wg/g)

Lindane,

2.6

2.0

Surface

appl

ed.

2.0

1.3

.42

Lindane,

10.4

.24

.27

6.5

12.2

Incorporated

9.1

.23

.21

8.6

9.6

6.8

.16

.13

5.7

6.1

Dieldrin,

100

.14

.12

189

109

Surface

appl

ed.

.12

176

.14

.08

126

Dieldrin,

10.2

.07

.02

Incorporated

9.4

.07

.02

8.3

.07

.01

1/ Unpublished data from A. W. White, 1970, Watkinsville, Ga.
Table 1-21. — Atrazine in runoff from simulated rainfall; comparison of observed— vs. predicted

Concentrati
runoff

on in

Concentration in
sediment

Percentage of
application

Size of Storm

Observed-

Predicted

Observed^' Predicted

Observed

Predicted

(cm)
1 hr
after
application

(mg/1)

(yg/g)

(ug/g)

(%)

1.25
3.18
6.35

7.2
2.3

1.3

2.2

1.9
1.7

24
9.4
4.6

4.3
12.0
17.0

1.4
10
24

96 hr
after
application

1.25
3.18
6.35

3.3
1.0
.55

1.3
1.1
1.0

11
4.2
2.0

7.8
6.6
6.0

2.0
5.3
7.3

.85
6.0
14

Enrichment ratio = 1.5.
Extraction ratio = 0.2.
Decay constant = 0.14.
Distribution coefficient, K^ = 4.

1/ White and others (15). Simulated rainfall, Cecil sandy loam soil,
2/ Estimated based on reported losses in water and sediment.

Table 1-22. — Atrazine in runoff; comparison of observed— vs. predicted;
Hagerstown silty clay loam, Pa.

Concentration in

Percentage of

Time after

runoff

sed

iment

appl'

ication

appl ication

Observed

Predicted

Observed

Predicted

Observed

Predicted

(days)

(mg/1)
4.6

(mg/1)

(mg/kg)
10

(mg/kg)
9.2

(%)
171

(%)
0.55

2.2

1.1

5.1

4.4

.70

.38

.92

.74

4.1

3.0

1.1

.90

.75

.34

3.3

1.4

.42

.19

.20

.24

1.8

1.0

1.1

.18

.08

1.7

.13

.10

.20

.05

2.2

.41

.08

TOTAL

4.9

3.3

cally.

Enrichment ratio = 2. Extraction ratio = 0.2. Decay constant = 0.05.
Distribution coefficient, K . = 2 (based on reported organic matter content
of 1.3%). a

1/ Hall, J. K. (7), observed values estimated from data reported graphi-

Table 1-23

-2,4-D in runoff from a semiarid desert rangeland
observedl' vs. predicted

comparison of

Concentration in

Time after

runoff

appl ication

Observed

Predicted

(days)

(mg/1)

0.240

0.26

.017

.064

.013

.054

.010

.018

.010

.012

.0067

.0064

.0029

.0015

.0042

.0011

Percentage of

appl ication

Observed

Predicted

(%)

0.424

0.454

.009

.035

.019

.074

.018

.032

.014

.017

.002

.018

.009

.004

.001

TOTAL

.508

.624

Enrichment ratio = (sediment concentration not computed). Extraction

ratio = 0.1. Decay constant = 0.1. Distribution coefficient, K

1/ Lane, L. J., 1978.

unpubl ished data.
99

Another set of comparisons on 2,4-D were obtained with the data of White
and others (14) (table 1-24). Here the soil was a loamy sand with high infil-
tration rates. Slope was about 3% and soil loss from the plots was relatively
low. However, the sediment phase was probably high in organic matter and clay
compared to the original soil, and consequently, the enrichment ratio would be
high. A value of 10 was arbitrarily assigned. Other parameters were estimated
from information in this report and observed 2,4-D persistence in the Cowarts
soil. The model gave good estimates of concentrations in the first storm, but
underestimated concentrations in the later storm. The greatest discrepancy
occurs in the comparison of concentrations in sediment. A K<j value of 1 was
assigned to 2,4-D based on information in volume III, chapter 19. However,
even with the assumption of a larger enrichment ratio, the observed partition-
ing would indicate an effective Kj of about 10 or greater which increased with
time. However, the objective of this evaluation is not to adjust parameters by
rationalizations to obtain fit to a particular data set. Since the 2,4-D form-
ulation used (alkanol amine salt) is very sparingly soluble in water, the K^
value of 1 used was probably inappropriate. However, using a higher Kj value
would not have greatly affected the predicted concentrations in the water
phase. Because of the large storms applied, predicted downward movement of the
pesticide significantly lowered the predicted runoff-available pesticide. Use
of a larger Kj would have prevented this, but in turn the portion entering the
water phase would be lowered by the higher Kj, and thus the predicted concen-
trations in solution would be similar. This illustrates how model predictions
can give good fit with observed total mass lost in runoff but not adequately
describe processes or "prove" that processes have been represented correctly.

The data on pesticide runoff reported by Smith and others (_n) provide an
opportunity to test the model in several ways. Tables 1-25 through 1-27 give
results for atrazine on a small watershed for three consequtive years. Tables
1-28 through 1-30 show results for three pesticides of differing properties on
another watershed in the same year, 1973. Many comparisons can be made; how-
ever, only a few points will be discussed. Predicted sediment concentrations
in the first storm agree reasonably well with observed. The agreement is very
poor for the later storms, indicating that assumptions regarding a constant
K<j are inadequate. It has been generally observed that pesticide desorption
from soil is nonlinear and appears to become more difficult with time. How-
ever, sufficient research information on pesticide behavior in soil is not
available for development of some model algorithm to account for this behavior
in a direct manner. Again, the objective here is not to calibrate some func-
tion to give close fit.

The data in tables 1-28 through 1-30 show that the model can give reason-
able predictions for pesticides of widely different properties and behavior.
However, the prediction for trifluralin is excessive. Since this compound is
quite volatile, the immediate soil surface subjected to runoff may have become
more depleted of pesticide than predicted in the model. In this particular ex-
periment, the data indicated that major transport was in the water phase.
Others (13) have indicated that sediment transport is the major route for this
compound.

Data in tables 1-20 through 1-30 are summarized in figures 1-22 through
1-25 for visual comparison. Observed vs predicted values are plotted in expo-
nential scale in figures 1-22 through 1-24 because of the range of values

100

Table 1-24. — 2,4-D in runoff obtained from a Cowarts loamy sand with simulated
rainfall of 8.25 cm in 30 min; comparison of observed!/ vs. predicted

Time after

Concentration in
runoff

Concentration
sediment

Percentage of
appl ication

appl ication

Observed

Predicted

Observed Predic

ted

Observed

Predicted

(days)

8
35

(mg/1)
0.022

.004

.0003

(mg/1)
07075

.002

< .0001

(mg/kg) (mg/kg)
2.8 6.25

1.5 .01

.6 < .001

(%)
1.53

.37

.06

(%)
1.54

.10

< .01

TOTAL

1.96

1.65

Enrichment ratio = 10. Extraction ratio = 0.1. Decay constant = 0.4,
Distribution coefficient, K, = 1.

1/ White, A. W. and others (14).

Table 1-25. — Atrazine in runoff; comparison of observed— vs. predicted water-
shed P2, 1973, Watkinsville, Ga.

Time after

Concentration in
water

Concentration in
sediment

Percentage of
appl ication

appl ication

Observed

Predicted

Observed

Predicted

Observed

Predicted

(days)
8

(mg/1)
0.200

(mg/1)
0.460

(mg/kg)
3.232

(mg/kg)
3.686

(%)

0.002

0.004

.179

.224

.856

1.793

.324

.421

.064

.094

.892

.752

.880

1.21

.044

.075

.785

.606

.528

.827

.014

.019

.879

.152

.118

.127

.017

.009

.749

.072

.008

.004

.036

.008

.900

.064

.006

.001

.007

.005

.425

.040

.021

.011

.009

.002

.587

.016

.015

.005

.002

< .001

.200

< .008

.015

< .005

TOTAL

1.92

1.61

Enrichment ratio = 2. Extraction ratio = 0.1. Decay constant = 0.14.
Distribution coefficient, K, = 4.

1/ Smith and others (11)

101

Table 1-26. — Atrazine in runoff; comparison of observed— vs. predicted, water-
shed P2, Watkinsville, Ga., 1974

Time after

Concentration in
runoff

Concentration in
sediment

Percentage of
application

appl i cat ion

Observed

Predicted

Observed

Predicted

Observed

Predicted

(days)
6

(mg/1)
1.90

(mg/1)
0.630

(mg/kg)
4.10

(mg/kg)
5.04

(%)

0.080

(%)

0.027

.022

.018

2.56

.144

.049

.035

.020

.0003

.910

.002

.008

< .001

.003

< .0001

.488

< .001

.012

< .001

.003

< .0001

.483

< .001

.032

< .001

TOTAL

.18

.06

Enrichment ratio = 2. Extraction ratio = 0.1. Decay constant = 0.14.

Distribution coefficient, K, = 4.

1/ Smith and others (11).

Table. 1-27. — Atrazine in runoff; comparison of observed— vs. predicted, water-
shed P2, Watkinsville, Ga., 1975

Time after

Concentration in
runoff

Concentration in
sediment

Percentage of
appl ication

appl ication

Observed

Predicted

Observed

Predicted

Observed

Predicted

(days)
10

(mq/1)
0.101

(mg/1)
0.122

(mg/kg)
1.53

(mg/kg)
0.979

(%)

0.784

(%)

0.T29

.017

.020

.987

.160

.098

.010

.017

.299

.136

.295

.402

.012

.005

.039

.040

.008

.003

.001

.0002

.038

.001

.006

.001

TOTAL

.69

.83

Enrichment Ratio = 2. Extraction Ratio = 0.1. Decay constant = 0.14.

Distribution coefficient, K
1/ Smith and others (11).

102

Table 1-28. — Trifluralin in runoff; comparison of observed— vs. predicted

Time after

Concentration in
runoff

Concentration in
sediment

Percentage of
appl ication

appl ication

Observed

Predicted

Observed

Predicted

Observed

Pred icted

(days)
0

(mg/1)
0.013

(mg/1)
0.037

(mg/kg)
0.0304

(mg/kg)
1.11

(%)

0.18

1.03

.0057

.012

.0321

.36

.024

.073

.0020

.0046

.0100

.138

.0011

.0036

.0045

.0011

.0206

.033

.021

.0063

.0050

.0003

.0574

.009

.0045

.00053

.0021

.00005

.0197

.0015

.0027

.0031

TOTAL

.233

1.12

Enrichment ratio = 2. Extraction ratio = 0.1. Decay constant = 0.14.
Distribution coefficient, K, = 15.

1/ Smith and others (11).

Table 1-29. — Paraquat in runoff; comparison of observed— vs. predicted

Time after

Concen
r

tration
unoff

Concenti
sed

"ation in
iment

Percent
appl ic

age of
ation

appl ication

Observed

Predicted

Observed

Predicted

Observed

Pred icted

(days)
0

(mg/1)
0

(mg/1)
1.5 x lO-4

(mg/kg)
36.8

(mg/kg)
30.4

9.70

8.02

1.4 x

10-4

35.6

28.7

1.36

1.09

1.4 x

10-4

61.5

27.2

.26

.12

1.3 x

10-4

29.7

25.5

.65

.56

1.2 x

10-4

38.0

24.0

.08

.05

1.1 X

10-4

27.6

21.9

1.75

1.38

TOTAL

13.80

11.22

Enrichment ratio = 2
Distribution coeffic

1/ Smith and others (11).

Extraction^ratio

Distribution coefficient, K , = 10 .

0.1. Decay constant = 0.007

103

Table 1-30. — Diphenamid in runoff; comparison of observed— vs. predicted

Time after

Concentration in
runoff

Concentration in
sediment

Percentage of
appl ication

appl ication

Observed

Predicted

Observed

Predicted

Observed

Predicted

(days)
0

(mg/1)
1.65

(mg/1)
1.90

(mg/kg)
0.64

(mg/kg)
3.80

6.82

(%)
8.42

.25

.21

.67

.54

.32

.35

.065

.077

.20

.15

.012

.014

.013

.008

.16

.016

.022

.012

.010

.001

.35

.002

.003

.0003

.002

.0001

.12

.0002

.01

.0004

TOTAL

7.19

8.80

Enrichment ratio = 2. Extraction ratio = 0.1. Decay constant = 0
Distribution coefficient, K, = 1.

1/ Smith and others (11).

.18,

obtained. The broken lines in each plot indicate 1:1 correspondence. Coeffi-
cients, r, shown are from linear correlation. Predictions of sediment concen-
trations in the first events were good except for trifluralin and diphenamid on
watershed PI (fig. 1-22). Acceptable predictions of solution concentrations
were obtained throughout most of the study periods (fig. 1-23 and 1-24). In
figure 1-24, all concentrations in excess of 1 ppb (0.001 ppm in tables) were
used in the comparison. Concentrations were somewhat underpredicted in late
events where they were very low. Use of different decay rates for different
times after application, as suggested as a possibility in volume III, chapter
17, could rectify this discrepancy. However, except for persistent pesticides,
runoff losses except during a short time period immediately following
application may be insignificant. Total losses for the growing season are
compared in figure 1-25.

In summary, the comparisons presented here demonstrate the potential of
the model and some of its shortcomings. Additional or other data sets probably
would have given different results, but most likely they would have fallen
within the range of accuracy indicated by these data.

Rather than use a comprehensive data set in an attempt to evaluate the
model in situations of multiple foliar applications of insecticide, hypotheti-
cal cases were set up that should closely resemble toxaphene application to
cotton as reported by Willis and others (_16j . It was assumed that six applica-
tions of 2.2 kg/ha were applied on days 2, 9, 16, 23, 30, and 37. Six rainfall
events of 2.0 cm each were imposed in two sequences: on days 3, 10, 17, 24,

104

31, and 38, and on days 8, 15, 22, 29, 36, and 43. The first sequence gave
rainfall 1 day after each application; the second sequence, 6 days after appli-
cation. Additionally, in each sequence above, it was assumed in one case that
10% of the application was intercepted by the soil with 50% intercepted by fo-
liage, and in another, none was intercepted by soil and 50% was intercepted by
foliage. The remaining pesticide was assumed off-target losses. An initial
toxaphene residue in the soil of 2 yg/g was also assumed. Parameter values and
inputs are summarized in table 1-31. Values of parameters were estimated based
on information provided by Willis and others (_16) and in volume III, chapter
18. Predicted toxaphene concentrations in water, sediment, and soil are shown
in figure 1-26 for three of the four scenarios. Concentrations increased in
all phases in response to application. However, when the insecticide
application was intercepted by foliage only, predicted concentrations

in6

o /

I05

I04

o /°

/ °

i«2

I01

/
/

R=0.96

,n<>

i i i

IOu I01 10* 10* 10* 10s

PREDICTED CONCENTRATIONS, ppb

Figure 1-22. — Comparison of predicted and

observed concentrations of pesticide in
sediment (first storm events after
application) .

105

increased little since only 10% washoff was allowed and the foliar half-life
was only 7 days. The 10% washoff may be greater than actually observed for
toxaphenes (see vol. Ill, ch. 18). The effects of delayed rainfall after ap-
plication are also apparent.

In the study by Willis and others (16), 10 kg/ha of toxaphene was applied
in six applications to cotton on a watershed containing about 2 ppm toxaphene
soil residue. Average toxaphene concentration observed in sediment during the
application period was 12.9 ppm. The maximum toxaphene sediment concentration
predicted here was 22 ppm after application of 13.4 kg/ha. Although no direct
comparisons can be made, model predictions appear to be reasonable compared
with observations. Total predicted losses in the six hypothetical storms are
summarized in table 1-32. These data illustrate the utility of the model in
examining relative effects of rainfall probabilities and pesticide application
efficiency. If the same simulations were performed with an insecticide of
shorter foliar half-life such as methyl parathion, the results would have been
affected more by timing than is reflected here. Since toxaphene with an esti-
mated Kd of 3500 is transported primarily by sediment, total losses also would
be sensitive to factors that reduce sediment yield.

io o

10 10

PREDICTED CONCENTRATIONS, ppb

Figure 1-23. — Comparison of predicted
and observed concentrations of
pesticides in solution phase of
runoff for first events after
appl ication.

0" 10' 10 10 10
EDICTED CONCENTRATIONS, ppb

Figure 1-24. — Comparison of observed
and predicted concentrations of
pesticide in solution phase of
runoff for all events with
concentrations greater than 1.0
ppb.

106

5 10 15
PREDICTED PERCENT

Figure 1-25. — Comparison of predicted and
observed seasonal losses of pesticide
in runoff, percent of total applied
or present during the year.

107

Table 1-31. — Parameters and inputs for simulation of hypothetical
cases of toxaphene appl ied to cotton

Appl ication dates
Appl ication rate
Rainfall dates

Days 2, 9, 16, 23, 30, 37

2.2 kg/ha

3, 10, 17, 24, 31, 38

8, 15, 22, 29, 36, 43

Rainfall amount

2.0 cm each event

Runoff amount

1.0 cm each event

Sediment yield

200 kg/ha each event

Depth of incorporation

1.0 cm

Incorporation efficiency

Fraction on fol iage

0.50

Fraction on soil

0.0
0.10

Foliar washoff rainfall
threshold.

0.10 cm

Initial fol iar residue

Initial soil residue

2.00 fig/q

Foliar washoff fraction

0.10

Water solubility

0.40 ppm

Foliar residue half-life

7 days

Enrichment ratio

1.50

Extraction ratio

0.10

Decay constant for
soil residue.

0.005

Distribution coefficient,

3,500

108

SOIL

«g*£\

SEDIMENT

WATER

20 30

DAYS

Figure 1-26. — Predicted toxaphene concentrations in 6
hypothetical storms of 2.0 cm, 1.0 cm runoff, and
200 kg/ha sediment yield: a = rainfall 6 days
after application (treatment 4, table 1-32); o =
rainfall 6 days after application (treatment 2,
table 1-32); and o= rainfall 1 day after appli-
cation (treatment 1, table 1-32).

109

Table 1-32. — Predicted losses of toxaphene in six hypothetical
storms of 2.0 cm each with 1.0 cm runoff and 200 kg/ha
sediment yield

Treatment- Mass in water Mass in sediment Total mass

(q/ha)
1.62

1.37

.74

.54

(q/ha)
17.20

(q/ha!
18.82

14.37

15.74

7.74

8.48

5.68

6.22

1/ (1) Soil residue + 6 applications at 2.2 kg/ha; 0.5 of
application on foliage; 0.1 on soil. Rainfall 1 day after ap-
plications (refers to n -symbols on figure 1-26); (2) Soil res-
idue + 6 applications at 2.2 kg/ha; 0.5 of application on foli-
age; 0.1 on soil. Rainfall 6 days after applications (refers to
o -symbols on figure 1-26); (3) Soil residue + 6 applications
at 2.2 kg/ha; 0.5 of application on foliage; 0 on soil. Rain-
fall 1 day after applications; (4) Soil residue + 6 applica-
tions at 2.2 kg/ha; 0.5 of application on foliage; 0 on soil.
Rainfall 6 days after applications (refers to a- symbol on fig-

ure 1-26

REFERENCES

(1) Baldwin, F. L., P. W. Santelmann, and J. M. Davidson.

1975. Movement of fluometuron across and through the soil. Journal of
Environmental Qual ity 4:191-194.

(2) Bruce, R. R., L. A. Harper, R. A. Leonard, W. M. Snyder, and A. W.
Thomas.

1975. A model for runoff of pesticides from small upland watersheds.
Journal of Environmental Qual ity 4:541-548.

(3) Crawford, N. H., and A. S. Donigian, Jr.

1974. Pesticide transport and runoff model for agricultural lands.
Hydrocomp, Inc., Palo Alto, Calif., prepared for U.S. Environmental
Protection Agency, Athens, Ga., Publication No. EPA-600/2-74-013.
211 pp.

(4) Davidson, J. M., G. H. Brusewitz, D. R. Baker, and A. L. Wood.

1975. Use of soil parameters for describing pesticide movement through
soils. U.S. Environmental Protection Agency, Publication No. USEPA-
660/2-75-009. 149 pp.

110

(5) Frere, M. H., C. A. Onstad, and H. N. Holtan.

1975. ACTMO, an aqricultural chemical transport model. U.S. Depart-
ment of Agriculture, Agricultural Research Service, Headquarters,
ARS-H-3, 54 pp. (Series discontinued; Agricultural Research Service
now Science and Education Administration-Agricultural Research.)

(6) Genuchten, M. Th . van, J. M. Davidson, and P. J. Wierengen.

1974. An evaluation of kinetic and equilibrium equations for the pre-
diction of pesticide movement through porous media. Soil Science
Society of America Proceedings 38:29-35.

(7) Hall, J. K.

1974. Erosional losses of _s-triazine herbicides. Journal of Environ-
mental Quality 3:174-180.

(8) Helling, C. S.

1971. Pesticide mobility in soils. II. Application of soil thin- layer
chromatography. Soil Science Society of America Proceedings 35:737-
743.

(9)

1971. Pesticide mobility in soils. III. Influence of soil properties.
Soil Science Society of America Proceedings 35:743-748.

(10) Leonard, R. A., G. W. Langdale, and W. G. Fleming.

1979. Herbicide runoff from upland Piedmont watersheds - Data and im-
plications for modeling pesticide transport. Journal of Environmen-
tal Quality 8:223-229.

(11) Smith, C. N., R. A. Leonard, G. W. Langdale, and G. W. Bailey.

1978. Transport of agricultural chemicals from small upland Piedmont
watersheds. U.S. Environmental Protection Agency, Athens, Ga. and U.
S. Department of Agriculture, Watkinsvil le, Ga. Final Report on In-
teragency Agreement No. D6-0381. Publication No. EPA 600/3-78-056.
363 pp.

(12) Steenhuis, T. S., and M. F. Walter.

1978. Closed form solution for pesticide loss in runoff water. Ameri-
can Society of Agricultural Engineers Technical Paper No. 78-2031,
presented at the 1978 summer meeting of the American Society of
Agricultural Engineers, Logan, Utah, June 27-30.

(13) Wauchope, R. D.

1978. The pesticide content of surface water draining from agricultur-
al fields - A review. Journal of Environmental Quality 7:459-472.

(14) White, A. W., L. E. Asmussen, E. W. Hauser, and J. W. Turnbull.

1976. Loss of 2,4-D in runoff for plots receiving rainfall and from a
small agricultural watershed. Journal of Environmental Quality 4:
487-490.

Ill

(15) White, A. W., A. P. Barnett, B. G. Wright, and J. H. Holladay.

1967. Atrazine losses from fallow land caused by runoff and erosion.
Environmental Science Technology 1:740-744.

(16) Willis, G. H., L. L. McDowell, J. F. Parr, and C. E. Murphree.

1976. Pesticide concentrations and yields in runoff and sediment from
a Mississippi Delta watershed. Proceeding of the 3d Federal
Interagency Sedimentation Conference, Denver, Colo.

112

Chapter 6. SENSITIVITY ANALYSIS
L.J. Lane and V. A. Ferreira—

INTRODUCTION

Sensitivity analysis is a technique for assessing the relative change in a
model response or output resulting from a change in inputs or in model parame-
ters. For simple, explicit models, it is possible to take derivatives of the
output with respect to input or parameters, and express the sensitivity as ex-
plicit functions. However, as the models become more complex, sensitivity is
more easily expressed in the form of differentials, relative changes, graphs,
and tables, rather than as functions. This is the approach used for the field-
scale model .

Based on derived parameter values and representative values of the input
variables, base values are selected. For a given set of base parameter values,
computations are performed, and then the input variables are varied over a
range of values and the computations repeated. For given values of the input
variables, the procedure is repeated with the parameters varying about their
base values. The resulting computations show how the model outputs vary with
changes in the input and parameters. This shows how the model functions and
how important each of the parameters is in determining the output. Such analy-
ses also aid in parameter estimation.

The main shortcomings of this procedure are (1) the parameters are varied
individually so that complex interactions are difficult to determine, and (2)
the number of simulation runs increases rapidly with the number of parameters
and inputs and with the number of points selected to vary about the base val-
ues. For example, nm + 1 simulation runs are required for a model with n pa-
rameters and input variables, and with simulation runs for the base values and
m points around the base value of each parameter and input variable. In some
cases, it may be necessary to limit the sensitivity analysis to a subset of the
model parameters. Finally, the sensitivity analyses given in this chapter are
for a complex watershed including detachment, transport, and deposition proces-
ses in overland flow and in concentrated flow. Sensitivity for other condi-
tions may be much different. Users should determine model sensitivity for the
particular application.

FIELD-SCALE MODEL: HYDROLOGIC COMPONENTS

The hydrologic components consist of two versions. The first, option 1,
uses daily rainfall to predict runoff volume and peak discharge rates. The

1/ Hydrologist and hydrologic technician, respectively, USDA-SEA-AR, South-
west Rangeland Watershed Research Center, Tucson, Ariz.

113

second, option 2, uses breakpoint precipitation data for individual events, and
also produces runoff volumes and peak discharge rates as output. Both options
also predict daily plant transpiration, potential transpiration, average soil
moisture, and percolation.

Option 1, Daily Rainfall Model

This model is essentially a modified Soil Conservation Service (SCS) run-
off curve number water balance model. As in the analysis for the other compo-
nents, data from watershed P2 at Watkinsville, Ga., were used to determine mod-
el sensitivity. This watershed and cultural practices on it are described in
detail by Smith and others (2J , and are summarized in table 1-33.

Table 1-33. — Summary of watershed characteristics and cultural practices for
watershed P2, Watkinsville, Ga., 1973-75

Item

Description

Area
Soil

3.19 acres

60% Cecil sandy loam

30% Cecil sandy clay loam

10% Loam

Cover

Corn, rows nearly on the contour

Cultural
' practices.

April 18, 1973
May 11, 1973

November 2
November 5

1973
1973

April 23, 1974
April 25, 1974
April 29, 1974

October 29, 1974

May 15, 1975
May 21, 1975

October 3, 1975

Tilled 20 cm deep with moldboard.—

Corn planted, 50,000 plants/ha, 90
cm rows.

Harvest.

Stalks shredded, 3,100 kg/ha stover,
Estimated near 30% residue cover.

Disked.

Chisel plowed, 20 cm deep.

Corn planted, 50,000 plants/ha, 90
cm rows.

Harvest, 6,300 kg/ha stover. Esti-
mated near 50% residue cover.

Disked.

Corn planted, 54,000 plants/ha, 90
cm rows.

Harvest, 6,800 kg/ha stover.

1/ See Smith and others (2J for additional practices, including fertilizer
and pesticide application rates and summary of hydrologic data.

114

Initial estimates of parameter values used in the sensitivity analysis
were made by J. R. Williams?/ as summarized in table 1-34. These parameter
values are denoted "base values," and the parameters were then varied about the
base values to determine model sensitivity.

Table 1-34. — Summary of model parameters for prediction of runoff volume and
peak discharge using the option 1 hydrologic model for watershed P2, Wat-
kinsville, Ga., 1974-75

Parameter Va?ue Comments

FUL 0.75 Portion of plant available water storage filled at
field capacity. Maximum value 1.0.

Soil evaporation parameter.

Runoff curve number for antecedent moisture condition
II. Selected using SCS National Engineering Handbook.

Channel slope determined from topographic map.

Plant available soil water storage in 7 layers of the
soil profile.

Leaf area index for corn throughout the growing sea-
son.

Watershed length-width ratio determined from topogra-
phic map.

Mean monthly air temperature.

Mean monthly radiation.

Note: Additional parameters not listed in this table were selected from
watershed characteristics using procedures outlined in volume II, chapter 1.

For 1974-75, 138 precipitation events were analyzed. Observed and pre-
dicted runoff volume and peak discharge for 48 runoff-producing events were re-
lated as follows:

^= 0.08 + 0.72 Q [1-211]

R2 = 0.53

C0NA

3.50

CN2

CHS

0.022

UL(I)

7 Values

X(D

9 Values

WLW

2.1

TEMP(I)

12 Values

RADI(I)

12 Values

and

^ = 0.39 + 0.77 Qp [1-212]

R2 = 0.30

2/ Hydraulic engineer, USDA-SEA-AR, Temple, Tex., personal communication.

115

where:

Q = observed runoff volume, in,

Q = predicted runoff volume using base values, in,

Qp = observed peak discharge, in/hr,

IJp = predicted peak discharge using base values, in/hr, and

R2 = coefficient of determination with 100 R2 as the percent variance
explained by the model.

Therefore, the model generally overpredicted runoff volumes for observed
volumes less than 0.30 in and underpredicted for larger values. The coeffici-
ent of determination for equation [1-211] is 0.53, meaning that the model ex-
plains 53% of the variance in runoff volume. For peak discharge, equation [I-
212], the model overpredicted for values of peak discharge less than 1.67 in/
hr, or virtually all but the very largest events. Also, the model only ex-
plained 30% of the variance in peak discharge (R2 = 0.30 in equation [1-212]).
From these results, it appears that the model predicts runoff volume better
than peak discharge. However, these results are for a specific watershed, and
may not be typical, especially for larger watersheds where daily rainfall may
be a better predictor for peak discharge. Finally, the results summarized by
equations [1-211] and [1-212] represent predictions rather than fitting or op-
timization results.

Based on these analyses, the base values of the parameters were judged as
adequate values to use in determining model sensitivity.

Sensitivity Analyses for Mean Runoff Volume and Peak Discharge

Mean values of predicted runoff volume and peak discharge were computed
for the 138 precipitation events for each run. As each parameter was varied
about the base value, the mean values were compared to the means from the "base
value" predictions as a measure of the sensitivity. The simulation data are
summarized in table 1-35. Column 2 of table 1-35 shows which parameters were
varied, and by how much. Columns 3 and 5 show the mean predicted runoff volume
and peak discharge, and columns 4 and 6 show the mean values divided by the
corresponding mean values predicted using the base values of all parameters.
Values of 1.0 in these two columns represent no change in runoff volume and
peak; values less than 1.0 represent decreases. For example, a 50% decrease in
the portion of plant-available water storage in the soil filled at field capa-
city, FUL, resulted in a 54% decrease in mean runoff volume and peak rate. A
33% increase in FUL resulted in over 100% increase in mean runoff volume, and
nearly 100% increase in the mean peak discharge. The results for FUL are sum-
marized in rows 2 to 5 of table 1-35.

As expected, there was an inverse relation between the evaporation parame-
ter, C0NA, and runoff as summarized in rows 6 to 9 of table 1-35. However,
runoff is more sensitive to decreases in C0NA than to increases. Runoff vol-
umes and peaks were more sensitive to the runoff curve number, CN2, than to any
other parameter. This is significant in that this parameter best reflects the
influence of management practices and crop cover. Moreover, estimation techni-
ques for this parameter are well developed in the SCS National Engineering
Handbook (3J •

116

Table 1-35. — Hydrologic model, option 1, sensitivity analyses, watershed P2,
Watkinsville, Ga., 1974-75 data; effect of parameter variation on runoff
volume and peak

Mean runoff

Mean peak

Run No.

Variation

Parameter

volume
Q

Q/Base

discharge
QP

Qp/Base

(1)

(2)

(3)

(4)

(5)

(6)

(%)

(in)

(in/hr)

Base

0.092

1.00

0.284

1.00

-50

.042

.457

.132

.465

-25

FUL

.059

.641

.185

.651

,+25
-7+33

.159

1.728

.474

1.669

.192

2.087

.565

1.989

-50

.143

1.554

.431

1.518

-25

C0NA

.130

1.413

.395

1.391

+25

.079

.859

.245

.863

+50

.073

.793

.228

.803

2/-10

.048

.522

.152

.535

-05

CN2

.067

.728

.210

.739

+05

.125

1.359

.379

1.335

+10

.172

1.870

.512

1.803

-50

.092

1.00

.254

.894

-25

CHS

.092

1.00

.271

.954

+25

.092

1.00

.294

1.035

+50

.092

1.00

.302

1.063

-50

.076

.826

.237

.835

-25

UL(I)

.084

.913

.261

.919

+25

.097

1.054

.299

1.053

+50

.101

1.098

.308

1.085

-50

.099

1.076

.304

1.070

-25

x(D

.096

1.043

.297

1.046

+25

.091

.989

.283

.996

+50

.090

.978

.279

.982

-50

.092

1.00

.323

1.137

-25

WLW

.092

1.00

.299

1.053

+25

.092

1.00

.272

.958

+50

.092

1.00

.263

.926

-50

.104

1.130

.321

1.130

-25

TEMP(I)

.095

1.033

.294

1.035

+25

.090

.978

.278

.979

+50

.089

.967

.274

.965

-50

.112

1.217

.341

1.201

-25

RADI(I)

.099

1.076

.305

1.074

+25

.088

.957

.273

.961

+50

.087

.946

.269

.947

1/ Upper limit for FUL is 1.0, a 33% increase over base value.
2/ Limits for CN2 are 0 to 100.

117

Fortunately, the model is less sensitive to plant-available soil water
storage, UL(I), leaf area index, X(I), average daily temperature, TEMP(I), and
average daily radiation, RAD 1(1). In terms of parameter estimation, this means
that rather coarse estimates of these parameters can be used with corres-
pondingly smaller errors in computed runoff. For example, monthly averages can
be used to interpolate daily values or point records can be used to represent
large geographic areas. This is important in applying the model to ungaged
watersheds.

Additional sensitivity analyses were conducted with respect to average
soil water content and mean volume of percolation. Table 1-36 shows the sensi-
tivity of soil water and percolation to leaf area index, temperature, and radi-
ation. In general, percolation was more sensitive than soil moisture with more
relative change in percolation with changes in temperature and radiation. How-
ever, a 50% error in temperature or radiation would represent very gross er-
rors. In a negative sense, these results suggest that the model is not very
sensitive to changes in crop growth and canopy, as reflected in the leaf area
index. A possible exception would be interactions, and thus changes in runoff
curve number with changes in crop canopy. Additional research may be needed to
determine interactions between runoff curve number and crop canopy development.

Table 1-36. — Effect of parameter variation on average soil water and percola-
tion

Mean

Run No.

Variation

Parameter

Mean SW

SW/Base

percol ation

Perc/Base

(%)

(in)

Base

0.353

1.0

0.137

1.0

-50

.361

1.02

.150

1.09

-25

X(D

.358

1.01

.141

1.03

+25

.352

.99

.135

.99

+50

.352

.99

.134

.98

-50

.364

1.03

.184

1.34

7
8

-25
+25

TEMP(I)

.357
.351

1.01
.99

.156
.127

1.14
.93

+50

.350

.99

.121

.88

-50

.368

1.04

.209

1.53

-25

RADI(I)

.360

1.02

.164

1.20

+25

.350

.99

.123

.90

+50

.349

.99

.112

.82

Model sensitivity for the hydrologic model, option 1, daily rainfall mod-
el, is summarized in table 1-37. As described at the bottom of table 1-37, a
parameter's sensitivity is judged as "significant" when changes in runoff due
to a parameter change exceed the absolute magnitude of the parameter change.
This criterion identifies areas where errors are "magnified" by the mode. For
this model and the particular data set analyzed, CN2, C0NA, and FUL are the

118

most sensitive parameters, with the runoff curve number, CN2, as the single
most important parameter. Therefore, the user must exercise good judgment and
particular care in selecting the runoff curve number.

Table 1-37. — Summary of sensitivity of daily runoff model for watershed P2,

Watkinsville, GaJ/

Parameter Mean volume Mean peak

Comments

CHS

None

Moderate

CN2

Significant

CONA

Significant

FUL

Significant

POROS

None

RAD 1(1)

Moderate

None

TEMP(I)

Moderate

UL(I)

Moderate

WLW

None

Moderate

X(D

Slight

Not considered in volume calculation,
peak coefficient is directly related to
CHS.

Critical parameter; small variation
causes gross change in runoff.

Strongly affects ET; increasing CONA
produces higher SW, inversely affects
runoff.

Portion of plant-available water stor-
age filled at field capacity; related
to soil properties.

Parameter in percolation calculation;
no effect on runoff.

Monthly average radiation.

Parameter in percolation calculation;
no effect on runoff.

Monthly average temperatures.

Plant-available soil water storage in
up to 7 soil layers.

Not considered in volume calculation;

influences peak discharge.

Monthly leaf area index; measure of
crop canopy.

1/ A + 50% change in parameters produces a change in mean runoff volume
or peak of: Slight < 10%; moderate 10-50%; significant > 50%. These sensitiv-
ity analyses are for a particular watershed and are thus site specific.

Option 2, Breakpoint Rainfall Model

This model uses breakpoint precipitation data as input to compute runoff
volume and peak rate, as well as soil moisture and percolation. As in the ana-
lysis described above, data from watershed P2, Watkinsville, Ga., were used to
determine model sensitivity.

119

Initial estimates of parameter values used in the sensitivity analysis
were made by R. E. Smiths/ as summarized in table 1-38. These parameter values
are denoted "base values," and then were varied about the base values to deter-
mine model sensitivity.

Table 1-38. — Summary of model parameters for prediction of runoff volume and
peak discharge using the option 2 hydrologic model for watershed P2, Wat-
kinsville, Ga., 1974-75

Parameter

Base value

Comments

FUL

0.75

CONA

3.5

2.0

26.0

13.0

RMN

.08

SLOPE

.025

XLP

350.

TEMP

(I)

12 Values

RAD I

(I)

12 Values

X (I)

7 Values

POROS

.41

.15

Same as in option 1.

Depth of surface soil layer (in).

Depth of maximum root growth layer

Effective capillary tension of soil

Manning's n for overland flow.

Slope of plane of watershed (ft/ft)

Length of plane (ft).

Same as in option 1

in).
(in)

For the period 1974-75, 138 precipitation events were analyzed. Observed
and predicted runoff volume and peak discharge for 58 runoff-producing events
were related as follows:

Q = 0.104 + 1.033 Q

[1-213]

and

R = 0.76

/6?= -0.060 + 2.084 Q

[1-214]

R = .75

where: Q = observed runoff volume, in,

1^ = predicted runoff volume using the base values, in,

Qp = observed peak discharge, in/hr,
1!jp = predicted peak discharge using base values, in/hr, and

R2 = coefficient of determination.

3/ Hydraulic engineer, USDA-SEA-AR, Fort Collins, Colo., personal communi-
cation.

120

Thus, the model generally overpredicted runoff volume and explained 76% of the
variance in runoff volume for the 3.2-acre test watershed. Runoff peak was al-
so generally overpredicted; the coefficent of variation is 0.75. As in the op-
tion 1 sensitivity test, base parameter values were chosen as they would be
chosen by the user, using available measurements and handbook values. No fit-
ting or optimization techniques were employed.

Sensitivity Analyses for Runoff Volume and Peak, Average Soil Moisture, and

Percolation

The effects of parameter variation on runoff volume and peak, average soil
moisture, and percolation were studied. As in the tests of option 1, the mean
of each predicted value was calculated (for 138 precipitation events) and com-
pared to the mean values obtained in the base run. The simulation data are
summarized in table 1-39. Columns 2, 4, 6, and 8 contain the mean predicted
values of each variable for each parameter variation. Columns 3, 5, 7, and 9
contain the ratios of the calculated means to their respective base values to
indicate the relative effect of each parameter variation. This relative effect
is compiled in table 1-40, which corresponds to table 1-37 of the option 1 sen-
sitivity study. Again, the results are for a particular watershed and relative
sensitivity may be different depending upon site specific conditions.

Model use objectives should be carefully determined before parameter val-
ues are chosen. For instance, if runoff volume prediction is the user's sole
interest, eight parameters affect this variable slightly, four moderately, and
only one, RC, significantly. Peak flow is not significantly affected by chan-
ges in any single parameter, but eight parameters have moderate influence;
their cumulative influence warrants some judicious choosing. Average soil
moisture is affected much like volume: slightly sensitive to seven parameters,
moderately sensitive to four, and significantly sensitive to two. Percolation
has been shown to be an extremely sensitive variable, responding significantly
to seven parameters, moderately to three, and slightly to three. Therefore, if
the model is to be used primarily for percolation estimates, seven of the para-
meters must be very carefully determined.

The parameters defining temperature, radiation, and leaf area index are
used in the calculation of evapotranspiration, and therefore, have greatest ef-
fect on percolation and soil moisture, and a resultant effect on runoff. Tem-
perature and radiation variation resulted in significant changes in percolation
and moderate changes in soil moisture and runoff volume. Because the parame-
ters are both directly related to ET, their effect on runoff, percolation, and
soil moisture is inverse, that is, increasing these parameters increases ET,
and thus decreases soil moisture, percolation, and runoff. The other ET para-
meter, CONA, is similarly related.

In option 2, a surface control layer, DS, is used to calculate runoff
volume. As the plant canopy develops (increased LAI), soil evaporation is
reduced proportional to exp(-0.4 LAI). At the same time, transpiration from
the entire soil profile may increase. The result of these interactions is that
moisture content in the surface control layer is not reduced as much as mois-
ture content in the entire soil profile. For specific storm sequences, the
runoff volume may be slightly increased (because of higher soil moisture in the
surface control layer) while the overall soil moisture in the entire soil pro-

121

Table 1-39. — Hydrologic model, option 2 sensitivity analysis, watershed P2, Watkins-
ville, Ga., 1974-75; effect of parameter variation on runoff volume and peak,
average soil moisture, and percolation

Mean Mean Mean avg.

Para- runoff peak soil Mean

Variation meter volume Q/Base disch. QP/Base moist. SW/Base perc Perc/Base
Q QP SW

(1) (2) 0) (4) (5) (6) (7) (8) (9)

(%) (in) (in/hr) Qn) (in)

Base 0.129 1.000 0.319 1.000 0.198 1.000 0.085 1.000

-50 .180 1.395 .398 1.245 .188 .946 .050 .582

-25 Dr .149 1.155 .354 1.109 .194 .980 .071 .829

+25 Kt .112 .866 .290 .908 .201 1.015 .098 1.149

+50 .101 .787 .276 .863 .204 1.026 .105 1.232

-50

.125

.973

.320

1.000

.110

.554

.129

1.511

-25

FUL

.126

.978

.311

.990

.155

.780

.107

1.258

+25

.131

1.021

.324

1.015

.246

1.241

.067

.784

+33

.132

1.026

.325

1.018

.263

1.325

.063

.735

-50 .167 1.297 .388 1.213 .102 .517 .121 1.420

-25 pnDfK .143 1.112 .349 1.092 .151 .762 .100 1.177

+25 KUKUi .119 .928 .305 .954 .254 1.278 .073 .854

+50 .111 .863 .289 .905 .315 1.588 .070 .826

"-50 ~120 ~942 ~3l5 ~971 ~207 IT542 Tll4 1~337~

-25 n<. .120 .959 .314 .983 .201 1.014 .095 1.118

+25 u:> .130 1.018 .323 1.011 .197 .995 .078 .911

+50 .130 1.033 .326 1.022 .197 .995 .073 .857

"50""" """129 1T06I ~321 i"664 ~208 lT546 Tl23 I~444~

-25 np .129 .999 .320 1.003 .202 1.018 .104 1.217

+25 w .129 1.004 .321 1.004 .200 1.006 .069 .813

+50 .129 1.005 .322 1.007 .201 1.015 .060 .707

-50

.167

1.300

.384

1.203

.191

.960

.059

.695

-25

.143

1.113

.348

1.089

.196

.986

.075

.880

+25

.115

.895

.294

.922

.201

1.012

.095

1.111

+50

.107

.833

.282

.883

.203

1.022

.100

1.172

-50

.128

.991

.258

.806

.199

1.001

.086

1.013

-25

SLOPE

.128

.997

.295

.922

.198

1.000

.086

1.005

+25

.129

1.000

.331

1.035

.198

1.000

.085

1.000

+50

.129

1.000

.341

1.067

.198

1.000

.085

1.000

-50

.171

1.329

.376

1.178

.237

1.192

.213

2.490

-25

CONA

.160

1.244

.360

1.128

.231

1.162

.195

2.288

+25

.121

.940

.309

.966

.191

.963

.068

.795

+50

.120

.929

.307

.962

.190

.960

.065

.763

-50

.130

1.010

.383

1.199

.198

.999

.084

.987

-25

RMN

.129

1.002

.351

1.100

.198

1.00

.085

.998

+ 25

.128

.994

.280

.878

.199

1.001

.086

1.009

+50

.127

.990

.249

.781

.199

1.001

.086

1.015

-50

.130

1.010

.383

1.199

.198

.999

.084

.987

-25

XLP

.129

1.002

.351

1.100

.198

1.000

.085

.998

+25

.128

.994

.280

.878

.199

1.001

.086

1.009

+50

.127

.990

.249

.781

.199

1.001

.086

1.015

-50

.127

.988

.318

.997

.245

1.237

.113

1.329

-25

X(D

.125

.970

.313

.979

.215

1.086

.095

1.112

+ 25

.130

1.006

.321

1.006

.193

.974

.082

.967

+50

.130

1.008

.323

1.012

.191

.962

.080

.938

-50

.155

1.206

.362

1.132

.262

1.320

.321

3.769

-25

TEMP(I)

.131

1.019

.324

1.015

.217

1.096

.105

1.232

+ 25

.127

.988

.317

.993

.188

.950

.074

.865

+50

.126

.981

.315

.987

.182

.920

.066

.773

-50

.141

1.094

.336

1.052

.270

1.359

.168

1.967

-25

RADI(I)

.133

1.034

.327

1.024

.231

1.163

.117

1.379

+25

.126

.981

.315

.987

.182

.919

.068

.797

+50

.125

.969

.314

.982

.172

.869

.055

.645

122

Table 1-40. — Summary of sensitivity of breakpoint runoff model for watershed

P2, Watkinsville, Ga.I/

Parameter

Mean
volume

Mean
peak

Mean
SW

Mean
perc

Comments

FUL Slight Slight Significant Significant

CONA Moderate Moderate Moderate

DS Slight Slight Slight

DP Slight Slight Slight

GA Moderate Moderate Slight

RMN Slight Moderate Slight

SLOPE Slight Moderate Slight

XLP Slight Moderate Slight

TEMP(I) Moderate Moderate Moderate

Significant
Moderate

Slight

Significant

RADI(I)Moderate Slight Moderate Significant

X(I) Slight Slight Moderate Moderate

POROS Moderate Moderate Significant Significant

RC Significant Moderate Slight Significant

Maximum value 1.0; im-
portant parameter in
soil moisture and
drainage.

Soil evaporation and
plant ET parameter.

Portion of soil profile
which defines initial
soil saturation.

Maximum root growth
layer (depth below DS).

Green and Ampt effec-
tive capillary tension;
used in infiltration
determination.

Manning's n for over-
land flow; affects peak
discharge.

Used in peak discharge
calculation.

Used in calculation of
ET; important to use
reasonable (measured if
possible) values.

Used in calculation of

ET.

Extremely important pa-
rameter; effects all
variables considerably.

Used in infiltration
determination.

1/ A + 50% change in parameters produces a change in predicted variable
of: Slight: < 10%; moderate: 10-50%; significant: > 50%.

123

file is decreased. Research is required for a better accounting of moisture
balance in the surface control layer. Without this improvement, slight in-
creases in runoff volume can occur with increases in leaf area index.

Though strongly influenced by the ET parameters, infiltration and drainage
are driven by the parameters DS, DP, GA, RC, POROS, and FUL, which define and
control the motion of water in the soil. DP and DS, soil depth parameters,
have minimal effect on runoff and soil moisture, but considerable influence on
percolation. GA and RC describe soil properties that govern the motion of
water into and in the soil. Their effect on average soil moisture is minimal,
but both have considerable effect on percolation (see table 1-39) and runoff.
Note that runoff is inversely related to both of these parameters. POROS and
FUL define soil capacities. Soil porosity significantly affects soil moisture
and percolation: greater porosity allowing greater quantities of water to be
stored, and thus less to be percolated. FUL limits the amount of soil moisture
that is available for plant use. It significantly affects percolation and soil
moisture, and has a slight resultant effect on runoff.

Runoff is affected by parameters that primarily drive ET and infiltration.
Both runoff volume and peak are directly affected by parameters that represent
watershed geomorphology and physical characteristics. Peak flow is a function
of flow volume and watershed characteristics. Because runoff peak is moderate-
ly sensitive to eight parameters, it is to be considered a sensitive variable.
If peak estimation is the primary purpose for model use, these eight parameters
must be carefully chosen. Fortunately, several are easily measurable or other-
wise determinable (for instance, good temperature and radiation data are avail-
able for many locations; watershed slope is often known, and leaf area index
can be reasonably estimated if crop type is known). The effective slope
length, XLP, and Manning's n, RMN, have identical moderate influence on predic-
ted peak discharge and slight resultant effect on volume of runoff, soil mois-
ture, and percolation.

FIELD-SCALE MODEL: EROSION/SEDIMENT YIELD COMPONENT

The main processes in the erosion/sediment yield component are overland
flow, concentrated flow, and impoundments (ponds). The overland flow component
uses a modified form of the Universal Soil Loss Equation (USLE) to compute sed-
iment detachment and the Yalin equation to compute sediment transport capacity.
A first-order equation is used to compute sediment deposition. The concentra-
ted flow component computes sediment transport and subsequent detachment or de-
position, depending upon flow conditions, using the Yalin sediment transport
equation, a flow detachment equation, and a deposition equation. The pond com-
ponent estimates how much of the sediment settles to the bottom of a pond be-
fore the flow passes through the impoundment.

Overland Flow Component

Selection of Parameters

The overland flow component has three input variables: EI = storm erosivi-
ty, Q = runoff volume, and ap = peak discharge rate. The USLE factors KCP all
occur in a linear form so that varying K shows the sensitivity to the other two
factors as well. The Manning's n value for surface-cover conditions, ncov> ex-

124

presses hydraulic roughness in overland flow and Cya] , the coefficient in the
Yalin sediment transport equation, represents the transport capacity. There-
fore, we chose to vary three input variables (EI, Q, ov, ) and three parameters
(K, ncov, Cya] ) to evaluate model sensitivity for overland flow.

Selection of Data and Base Values

Data from watershed P2 at Watkinsville, Ga., were also used to study model
sensitivity for the overland and concentrated flow components. These data are
summarized in table 1-33.

Base values for many of the parameters are summarized in table 1-41. Ob-
served values of EI, Q, and p for 32 storm events were used to simulate sedi-
ment yield for the 1973-74 period. Estimated sediment yield from the watershed

Table 1-41. — Summary of model parameters selected for prediction of sediment
yield from watershed P2, Watkinsville, Ga., 1973-75

Parameter

Base value

Comments

K 0.23

C 0.11-0.68
P 1.0

n 0.010-0.035
cov

Sal °-635

Soil erodibility factor; does not con-
sider seasonal variability (see vol. II,
ch. 2.)

Soil loss ratio; reflects cover and cul-
tural practices (see vol. II, ch. 2.)

Contouring factor; up and downhill till-
age assumed (see vol. II, ch. 2.)

Manning's n for overland flow; reflects
effect of cover and cultural practices
on flow's hydraulic resistance.

Empirical constant in Yalin sediment
transport equation; published value
used (see vol . II , ch. 2.)

K , 0.135

rch

Soil erodibility factor for erosion by
concentrated flow; value empirically de-
rived from rill erosion studies.

T 0.15-0.50

Critical shear stress for erosion by
concentrated flow; reflects effect of
cover and cultural practices on resis-
tivity of soil to detachment by concen-
trated flow.

nch 0.03-0.12

'bch

0.03

Manning's n for concentrated flow;
reflects effect of cover and cultural
practice on flow's hydraulic resistance.

Manning's n for concentrated flow over
bare, tilled agricultural soil; value
empirically determined.

Note: Parameters, including those not listed in this table, were selected
using procedures outlined in the user Manual, vol. II, ch. 2.

125

2.0

2 '•»

cr
o

i 1.0

0.5

1 1 1 1

WATKINSVILLE, GA.

WATERSHED P2

SELECTED DATA 1973-75

PREDICTIONS USING OBSERVED
RUNOFF DATA

^<^ -

o
o

^S*\ — Y = 0.09 + 0.46X
° ^^ 2

^^ R =0.63

^*^ o

0 ^^^

w°

I 1 1 1

0 0.5 1.0 1.5 2.0 2.5
OBSERVED SEDIMENT YIELD (TONS/ACRE)

Figure 1-27. — Relation between observed and pre-
dicted sediment yield for selected storms
on watershed P2, Watkinsville, Ga.

was used as a model output, and these yields were then summed for the 3 years
to produce a total sediment yield in tons per acre. The results of these ini-
tial predictions or base value runs are shown in figure 1-27. The model under-
estimated sediment yield for three largest events. The model explained over
60% of the variance in observed sediment yield. The total observed sediment
yield was 8.26 tons/acre, and the computed sediment yield was 6.56 tons/acre,
with a ratio of estimated-to-observed of 0.79 representing a 20% error in total
sediment yield. However, the 95% confidence limits for the mean observed sedi-
ment yield per event are 0.258 + 0.188 or 0.07 to 0.45 tons/acre. The mean
predicted sediment yield was 0.205 tons per/acre, well within the confidence
1 imits.

Sensitivity Analysis for Overland Flow Component

Computed sediment yields for the sensitivity analysis are summarized in
table 1-42. Column 2 shows which item was varied and that each was varied over
+ 25 and + 50% of the base value. Column 3 shows the computed sediment yield
from overland flow, and column 5 shows the ratio of this yield to the yield for
the base values. Column 7 shows the ratio of overland to total watershed sedi-
ment yield and is similar to a delivery ratio. Except for run no. 21, all sim-
ulations indicate net deposition in. the channel system, so that sediment yields
from the watershed were less than sediment yields in overland flow. However,
for 5 of the 32 events the model predicted net erosion in the channel system
producing a delivery ratio greater than one. For the other 27 events, the mod-
el predicted net deposition in the channel system producing an overall effect
of less sediment yield from the watershed than from overland flow.

Sensitivity of the model output to changes in EI, Q, and ap for the 32

126

Table 1-42. — Overland flow sensitivity analysis, watershed P2, Watkinsville,

Ga., selected data, 1973-75

Ratio of

Run

Overland

Watershed

watershed to

no.
(1)

Variation
(2)

sed. yield
QSO

(3)

sed. yield
QSW

(4)

QS0/Baseo
(5)

QSW/Basew
(6)

overland

sed. yield

QSW/QSO

(7)

(%)
Base

(tons/acre)
8.700

(tons/acre)
6.560

1.00

0.754

-50

6.685

5.186

.768

.791

.776

-25

7.677

5.823

.882

.888

.758

+25

9.481

7.004

1.131

1.068

.739

+50

10.155

7.605

1.167

1.159

.749

-50

5.603

4.178

.644

.637

.746

-25

7.252

5,375

.834

.819

.741

+25

9.810

7.420

1.128

1.131

.756

+50

11.004

8.432

1.265

1.285

.766

-50

6.136

4.773

.705

.728

.778

-25

<?n

7.616

5.749

.875

.755

+25

9.376

6.953

1.078

1.060

.742

+50

9.930

7.373

1.141

1.124

.742

-50

5.365

4.391

.617

.669

.818

-25

7.281

5.600

.829

.854

.769

+25

9.905

7.352

1.139

1.121

.742

+50

10.634

7.840

1.222

1.195

.737

-50

13.784

7.650

1.584

1.166

.555

-25

11.161

7.154

1.283

1.091

.641

+25

ncov

6.505

5.672

.748

.865

.872

+50

4.978

5.043

.572

.769

1.013

-50

5.977

4.666

.687

.711

.781

-25

Sal

7.623

5.821

.876

.887

.764

+25

9.427

7.057

1.084

1.076

.749

+50

10.029

7.527

1.153

1.147

.751

storms is shown in figure 1-28. Overall, the percent change in sediment yield
was approximately half of the percentage change in input variables. The model
was more sensitive to decreases in the input variables than to increases. Ex-
cept for a large increase in EI, changes in sediment yield were nearly linear
with changes in input variables, and the model appears to be most sensitive to
volume of runoff, Q. »

Overall, the percent change in sediment yield was approximately half the
percent change in K and Cya] , the soil erodibility and Yalin transport equation
coefficient. Again, the relative changes were greater for decreases in the pa-
rameters than for increases. For the hydraulic roughness, Manning's ncov, the
changes in sediment yield were larger than the relative changes in the parame-

127

? o

Q -1

z o

Z Q

< -I

I UJ

-25-

-50

WATKINSVILLE, GA.
WATERSHED P2
SELECTED DATA 1973-75

-50 -25 0 25 50
% CHANGE IN INPUT VARIABLES

Figure 1-28. — Sensitivity of sediment yield in
overland flow to input variables.

ter. This is significant in that this roughness parameter has a direct control
on transport capacity, and thus on deposition in overland flow. The change in
overland flow sediment yield appeared to vary nearly linearly with changes in
the roughness parameter over the range that the parameter was varied. For the
observed data on this particular watershed, the computed overland flow sediment
yield was most sensitive to this parameter.

Interpretations and Summary

Typical slope values for watershed P2 varied from 2 to 6% for overland
flow and 1.5 to 3.5% for the channel or concentrated flow section. Changes in
sediment yield with changes in roughness indicates deposition in overland flow.
As discussed earlier, the channel component produced a "delivery ratio" of less
than one. Sediment yield was limited by transport capacity for many of the 32
storms, and was primarily controlled by deposition. These sensitivity analyses
represent results obtained for a watershed situation where transport capacity
was moderately limiting for most storms. These results may be typical for cul-
tivated fields with low to moderate concave slopes and with relatively high
values of hydraulic roughness due to tillage and residue cover conditions. In
this situation, the hydraulic roughness factor for overland flow should be
carefully chosen.

Sensitivity Analysis for Concentrated Flow Component

Sensitivity analyses for the concentrated flow (channel) component can be
separated into two sections: (1) inputs and parameters affecting overland flow
sediment delivery to the channel system, and (2) parameters directly affecting
detachment, transport, and deposition in the concentrated flow.

128

Overland Flow Factors Affecting Sediment Delivery to the Channel

Column 4 in table 1-42 shows total sediment yield from the entire water-
shed, and the difference between these values and corresponding values listed
in column 3 show the net effect of the channel on sediment yield.

In all cases except run no. 21, the net effect of the channel was a sink
rather than a source for the total sediment yield from the 32 storms. Some of
the larger events showed net channel erosion, but the overall effect was to re-
duce sediment yield. Also, erosion was often predicted near the upper end of
the channel, with deposition predicted near the outlet where water was ponded
by a runoff measuring flume.

The influence of changes in EI, Q, and <rp on sediment yield from the wa-
tershed is shown in table 1-42. The results for EI and a^ are very similar to
those in overland flow only (figure 1-28), but the influence of runoff volume,
Q, is more pronounced, suggesting that this may be a most sensitive input. In-
fluence of the "overland" parameters (K and ncov) are shown in table 1-42.
Note that the influence of overland flow hydraulic roughness is damped out by
the channel system so that its sensitivity is comparable to the other para-
meters.

Sediment routing was extended to the entire watershed system with the re-
sult that the influence of runoff volume was accentuated and the influence of
overland flow parameters was dampened.

Parameters Directly Affecting Sediment Yield from Concentrated Flow

Channel routing assumptions—The general case for concentrated flow in a field
situation is a channel of length L with an upstream inflow rate, Q-j , and a la-
a lateral inflow rate, q* , along the channel reach. This configuration is il-
lustrated in figure 1-29, with the runoff rates corresponding to the peak dis-
charge at steady state, that is, steady-state spatially varied flow with in-
creasing discharge. The effective channel length, Leff, is the length of chan-
nel required to produce the outflow discharge, Qe given the lateral inflow
rate. The procedure used here is to solve the spatially varied flow equations
for a channel of length Leff to produce depth, velocity, and shear stress along
the channel reach, and then apply the transport and detachment capacity
equations along the original length of channel, L, to compute sediment yield
for the channel .

Assuming a wide range of channel and flow conditions, the spatially varied
flow equations were solved, and polynomial equations were fitted by regression
to the solutions (vol. I, ch. 3). As part of these sensitivity analyses, the
regression equations were reviewed and found to be quite accurate over a wide
range of conditions representing subcritical flow in channels with triangular
cross-sections. Although the model has user options for rectangular and natur-
ally eroded channels, the regression equations for spatially varied flow have
not been checked under these conditions. Therefore, results of the sensitivity
analysis presented here are for triangular channels only. A second user option
is to assume the friction slope in the routing procedure to be equal to the
channel slope. This assumption has been tested under a limited number of

129

Figure 1-29. — Illustration of general case for concentra-
ted flow in a field-sized channel: (A) schematic of
watershed channel system, (B) channel reach with up-
stream inflow and uniform lateral inflow, and (C) ef-
fective channel reach for spatially varied flow com-
putations.

conditions and has been found to be a poor approximation except under special
circumstances. The conditions under which this assumption is appropriate are
(1) no outlet controls producing backwater; (2) channel slopes relatively
steep; and (3) lateral inflow rate per unit length of channel is small compared
with the outlet discharge.

In summary, the model assumes a triangular shaped channel to estimate
friction slopes. The same friction slope estimates are used for rectangular
channels. Preliminary analyses suggested that this approximation can be quite
accurate. However, additional tests under a variety of conditions are required
before the general applicability of the procedure can be determined.

130

Channel erosion/sediment transport parameters—The parameters selected for
analysis here are summarized in table 1-41. In addition, selected user options
were used to determine model sensitivity to these inputs. Simulation runs for
the concentrated flow sensitivity analysis are summarized in table 1-43. Most
items were varied over + 25 and + 50% of their base values except for runs 34
to 41, where Manning's n for the cover-practice was restricted to be equal to
or larger than Manning's n for bare soil in the channel. Column 4 shows the
computed sediment yield from the entire watershed; column 5 shows the ratio of
this yield to the yield using the base values. Column 6 shows the ratio of
overland to total watershed sediment yield and is similar to a delivery ratio
for the simulation results. In all runs except 44 and 45, simulations suggest
that the channel system was a "sink," so that sediment yields from the water-
shed were less than sediment yields from overland flow. For a few events, net
channel erosion was estimated, but the overall effect was net deposition in the
channel .

Sensitivity of the model output to changes in Krcn, rcr, and ncn for the
32 storms is shown in table 1-43. For the erodibility factor, Krcn, and criti-
cal shear stress, rcr, the model was more sensitive to decreases than to in-
creases in the parameters. For hydraulic roughness, ncn, the situation was re-
versed in part due to the constraint that ncn >_ nDCn. However, total sediment
yield was more sensitive to the critical shear stress than to the erodibility
or roughness parameters.

Changes in sediment yield were nearly linear with the side slope, but were
nonlinear and very large with changes in the channel slope. Although a + 50%
change in channel slope is an extreme error in user input, nonetheless, the
simulated sediment yield is very sensitive to channel slope.

The influence of assuming a rectangular or naturally eroded channel cross-
section was a 30% increase in estimated sediment yields over similar predic-
tions using a triangular channel. The most significant increase in estimated
sediment yield (+88%) was due to assuming that the friction slope was equal to
the channel slope. The reason for this is that the H-flume measuring structure
used on this watershed caused significant backwater, which is ignored when
using the channel slope as an approximation to the friction slope. This is a
most serious error, yet a common one in runoff and sediment routing, and re-
quires a good deal of judgment by the user. The fact that the simulated sedi-
ment yield was more sensitive to the "channel slope" assumption than to + 50%
errors in input variables or parameters suggests that the user exercise caution
in specifying the outlet control. Site specific conditions such as grass
around field edges or other outlet controls may pond the water and induce sig-
nificant deposition.

Interpretations and Summary

Model simulations and limited field observations indicate that sediment
yield from watershed P2 is primarily transport limited, and thus is primarily
controlled by deposition. However, for certain parameter values and for the
largest storms, the model predicted significant channel erosion. From these
results we may conclude that the concept of a constant delivery ratio is, at
best, a gross approximation.

131

Table 1-43. — Concentrated flow sensitivity analysis, watershed P2, Watkins-
ville, Ga., selected data 1973-75

Ratio of

Watershed

watershed to

Run no.

Variation

Parameter

sed. yield
QSW

QSW/Basew

overland

sed. yield

QSW/QSO

(1)

(2)

(3)

(4)

(5)

(6)

(%)

(tons/acre)

-50

^rch

6.261

0.954

0.720

-25

6.414

.978

.737

+25

6.696

1.021

.770

+50

6.828

1.041

.785

-50

rcr

7.911

1.206

.909

-25

7.184

1.095

.826

+25

6.204

.946

.713

+50

6.016

.917

.691

-50

nch

1/6.753

1.029

.776

-25

6.635

1.011

.763

+25

6.064

.924

.697

+50

5.676

.865

.652

-50

nbch

1/5.546

.996

.637

-25

5.558

.998

.639

+25

5.574

1.001

.641

+50

5.601

1.006

.644

-50

3.520

.537

.405

-25

Slope of

4.138

.631

.476

+25

Channel

9.269

1.413

1.065

+50

11.788

1.797

1.355

-50

7.697

1.173

.885

-25

Side

6.994

1.066

.804

+25

Slope of

6.262

.955

.720

+50

Channel

6.048

.922

.695

1/ Lowest values of ncn set to nbcn = 0*03 to insure ncn >_ nbch*
2/ Lowest values of ncn set to 0.045 to insure that ncn >_ nbch •
fore, the base value for these 4 runs was QSW = 5.568 tons/acre.

There-

The common "normal flow" or "kinematic assumption" of friction slope equal
to the bed slope can lead to serious errors in cases where an outlet control
causes significant backwater effects. Therefore, in applying the model, site
specific conditions causing possible backwater effects should be carefully
evaluated to determine the proper outlet control (discharge-depth relation-
ship).

132

Model sensitivity for the overland and concentrated flow components is
summarized in table 1-44. In this situation, overall watershed sediment yield
was less than overall overland flow sediment yield, suggesting a transport lim-
iting or depositional channel. For this reason, sediment yield was more sensi-
tive to "transport-deposition" inputs and less sensitive to "erosion-detach-
ment" inputs. In steep watersheds with actively eroding channels without out-
let controls, this situation can be reversed. In either case, site-specific
conditions such as outlet controls on channel depth-discharge relationships can
be significant in determining sediment yield.

Impoundment (Pond) Component

The impoundment or pond component was tested using data from three small
watersheds in Iowa with impoundment terraces {1 ) . The data are from watersheds
at Charles City, Eldora, and Guthrie Center, Iowa, with drainage areas of 4.6,
1.8, and 1.4 acres, respectively. Sediment yield data were measured at the
outlets with overland flow runoff volumes and peak rates used to estimate the
overland flow sediment input to the ponds. The impoundment component is based
on the rate particles reach the bottom of the impoundment versus the rate that
they leave the impoundment with the outflow. With this simplified model, the
fraction of particles of a specific size and density that passes through the
pond follows an exponential distribution.

Observed and computed sediment yield data from the three impoundment ter-
races are shown in table 1-45. Relations between observed and computed sedi-
ment yields is shown in figure 1-30. In general, sediment yields were underes-
timated a, as shown by the regression equations in figure 1-30. However, the
simulation results, especially those for Guthrie Center, were judged adequate
to use as base values in determining model sensitivity.

Sensitivity of the model to the infiltration rate into the pond bottom is
summarized in table 1-46. In this analysis the volume of water discharged from
the pond was held constant. Therefore, as infiltration in the pond was in-
creased from 0 to 0.5 in/hr the volume of runoff reaching the pond had to be
increased to maintain the same outflow. The estimated sediment yield in over-
land flow increased two to three times with the increase in volume of runoff.
Simulated sediment yield from the ponds increased only 1 to 40%. Only 3 to 11%
of the input sediment passed through the ponds. Although the 1 to 40% changes
in sediment yield at the impoundment outlets are significant, they are an order
of magnitude, or more, less than the corresponding changes in overland flow
sediment yield. However, as expected, there was significant clay/silt
enrichment with a much higher proportion of the sediment leaving the
impoundment in the clay-size range.

Increasing the soil erodibility factor K by 50% resulted in sediment yield
increases from the impoundments of 45, 45, and 18% for the three watersheds,
respectively. These changes are of the same order of magnitude as the increas-
ed sediment yield in overland flow. This suggests that the impoundment compo-
nent is sensitive to the sediment load entering the impoundment and its parti-
cle-size distribution.

133

Table 1-44. — Summary of sensitivity of erosion sediment yield model for water-
shed P2, Hatkinsville, Ga.1/

Input

KCP

-yal

cov

Type

Relative sensitivity of
total sediment yield

Comments

Variable
Variable

Variable

Parameter

Parameter
Parameter

Krch

Parameter

rcr

Parameter

nch

Parameter

nbch

Parameter

Channel

User

shape.

option

Friction

User

slope.

option

Moderate!/
Moderate

Moderate

Moderate
Moderate

Significant

Moderate

Slight

Significant

Moderate

Significant

Measure of detachment capacity of
rainfall .

Measure of detachment and transport
capacity of overland and concentra-
ted flow.

Overland flow detachment parame-
ters.

Transport parameter.

Overland flow hydraulic roughness.
Less sensitivity for watershed sed-
iment yield with depositional chan-
nels.

Not sensitive for depositional
channel .

Sensitive for depositional channel.

Channel slope should be carefully
determined.

Triangular
slope.

Site specific

cross-section

side

Most critical user option. Outlet
control should be carefully deter-
mined.

1/ Watershed with "depositional " channel on the average.

2/ A + 50% change in input produces a change in sediment yield of: Slight
- 0 to 10%; moderate - 10 to 50%; significant - greater than 50%.

3/ These results are for a particular watershed and a given sequence of
storms. Relative sensitivity of the model to input parameters can be different
under different conditions.

134

Table 1-45. — Summary of observed and simulated sediment yield from impoundment

terraces in Iowa

Water;

shed

Jul ian
date

Observed
sediment yield

Computed
sediment yield

Charles

City

70147

(lb)
1,197

(lb)
52

70152

70244

160

70323

71151

280

294

71157

209

160

Eldora

68198

283

150

68220

69187

1,057

554

69232

124

227

71163

335

139

Guthrie

Center

69207

256

273

69249

70144

122

70162

198

123

70167

70229

The assumed particle size distribution for the eroded soil is shown in ta-
ble 1-47. Two alternate distributions with smaller clay and smaller clay and
silt particles are also summarized in table 1-47. Smaller particles should
have a lower fall velocity, and thus more of them should pass through an im-
poundment. These simulation results are summarized in table 1-48. Decreasing
the diameter of clay particles from 0.002 to 0.001 mm resulted in an average
18% increase in sediment yield from the impoundments. With the same reduction
in clay size and reducing the diameter of silt particles from 0.010 to 0.005 mm
resulted in an average 56% increase in sediment yield from the impoundments
(table 1-48). This means that the model predicts significant clay and silt en-
richment, and that the impoundment model is quite sensitive to the assumed par-
ticle size distribution. Therefore, for depositional systems such as impound-
ments, accurate particle-size data are critical.

135

400

NO LINEAR RELATION
Q8=I25 + 0.3QS

R2-O.OI

CHARLES CITY

800

1200

GUTHRIE CENTER

0 400 800 1200
OBSERVED SEDIMENT YIELD (LB)

Figure 1-30. — Relation between observed and
computed sediment yield from impoundment
terrace outlets in Iowa.

136

Table 1-46. — Summary of simulated sediment yield from impoundment terraces in
Iowa with the infiltration rate in the pond assumed to be 0 , 0.2, and 0.5
in/hr

hed,
date

Sediment yields

in overland

flow from

pond outlet

Waters
Jul ian

I = 0.0

in/hr

I = 0.2

in/hr

I = 0.5

in/hr

Overland

Pond

Overland

Pond

Overland

Pond

Charles C

ity

70147

1,257

1,361

1,824

70152

150

219

315

70244

346

151

420

155

585

160

70323

110

71151

2,775

326

6,005

323

11,571

294

71157

1,824

125

3,580

121

6,176

160

TOTAL

6,462

681

11,657

677

20,543

685

drI/

—

0.11

0.06

—

0.03

Eldora

68198

1,714

134

2,559

140

4,612

150

68220

646

1393

1,577

69187

8,091

490

13,495

413

27,122

554

69232

2,250

178

4,982

189

7,172

227

71163

1,349

2,296

4,222

139

TOTAL

14,050

933

24,725

879

44,660

1,125

0.07

0.04

0.03

Guthrie Center

69207

,849

192

4,999

241

5,935

273

69249

825

1,104

1,844

70144

356

413

1,301

70162

578

112

1,149

120

1,704

123

70167

121

195

390

70299

113

258

451

TOTAL

,842

449

8,118

518

11,625

628

—

0.05

1/ DR is the delivery ratio; sediment yield from the pond outlet divided
by the sediment yield in overland flow.

137

+j o

<T3 -i-

O >

QJ A3
CL 1_
CO en

OJ C

+J O

ro ••-

C 4->

i- 13

+J -i-

-— J-

< +->

u >

<D «3
O. S-
CO CH

-O T-

<D +J

•r- >,

E ^

<4- 4->

3 .O

CO T-

U >

CO S_

OJ rt3

< 4->

CL t_

CO CD

.^oJ

-t-> Q.

-0

i- >>

fO +J

1— 1

o o

CO 00

O CM

oo|

E O

U CO
CD CM

i— l <— I CM

o o

CO CO

O CM

en c
en o

fO -r-
4->

■M E

r- 3

•r- CO

CO CO

>> E

(T3 O
i— i_
U "4-

CO u
QJ

CL-C

|Cvj|

138

Table 1-48. — Influence of particle-size distribution on sediment yield from im-
poundment terrace outlets

Watershed

Total sediment yield from impoundment outlet

Assumed Alternate Alternate
distribution distribution 1 distribution 2

Charles City
Eldora
Guthrie Center

(lb;

677
879
518

(lb)

908

961
581

(lb)
1,085

1,517

702

SUMMARY

We emphasize that the results of this sensitivity analysis are very site
specific and also specific to the observed storm sequences. Therefore, the re-
sults are indicative of a particular application and do not necessarily apply
in general. The model user should conduct sensitivity analysis for his speci-
fic conditions.

The Watkinsville watershed is a mixed, complex watershed. By that we mean
the control of sediment yield was mixed between detachment and deposition, be-
tween overland flow and channel flow, between storm sizes and sequences, and
between particle size classes. While primarily deposition controlled, detach-
ment had a significant effect. For some storms, little deposition occurred
either in overland flow or in channel flow. Often the upper ends of the over-
land and channel flow areas eroded while sediment was deposited in the lower
ends. Furthermore, for some storms the net effect for the coarse particles was
deposition while the net effect for fine particles was erosion. Consequently,
sediment yield was sensitive to both detachment and transport parameters.

In more simple situations such as uniform overland flow slopes or uniform
grade channels, many of these complex interactions would not occur. This would
be true also for analysis of single storm events. For example, in overland
flow where transport capacity exceeds incoming sediment load, the sediment
yield may be controlled by detachment parameters. As a result, sediment yield
may not be sensitive to Manning's n. However, with increasing roughness as the
transport capacity becomes controlling, the sediment yield would be more sensi-
tive to changes in Manning's n. Again, whether sediment yield is detachment or
transport limited, will depend upon site specific conditions and the rates and
amounts of runoff. Similar analysis can be made for a channel. In any event,
it is necessary to consider simple flow systems to isolate and study the beha-
vior of model components with changes in parameter values. For this reason, we
recommend that these sensitivity analyses be considered as examples of a spe-
cific application and that the model user conduct similar analyses for his par-
ticular appl ication.

139

FIELD-SCALE MODEL: CHEMISTRY COMPONENT

The chemistry component of the model consists of a nutrient submodel to
account for plant nutrients and pesticide submodel to account for pesticides.
The hydrologic component provides input to the erosion/ sediment yield compo-
nent which in turn provides input to the chemistry component. The input to the
nutrient submodel consists of rainfall, runoff, and sediment yield as well as
climatic variables necessary to simulate a water balance including runoff,
evapotranspiration, soil moisture, and percolation. Since the water balance
calculations are not as critical in the pesticide processes, primary inputs to
the pesticide submodel consist of rainfall, runoff, sediment yield, and an en-
richment factor. In these analyses, we used observed rainfall, runoff, and
sediment yield as input to the chemistry component.

Nutrient Submodel

The nutrient submodel is an accounting and transport model to estimate ni-
trogen and phosphorus losses from fields. Nutrients are added to the system as
fertilizer; in addition, nitrogen is added by rainfall and by mineralization of
organic matter from crop residue. Chemical transport is in the solution and
sediment phases. Nitrate leaching, plant uptake, and dentrification are calcu-
lated to complete the mass balance. Analyses in this section are limited to
option 1, wherein the amount of dry matter is estimated from the yield poten-
tial and the ratio of actual transpiration to potential transpiration. Using
this procedure, the fraction of the total plant growth expected is calculated
and the amount of nitrogen currently in the plant material is then the product
of the dry matter and average concentration in the dry matter. Incremental
plant uptake of nitrogen is then the difference between the current and pre-
vious value of nitrogen in the plants.

Output from the nutrient submodel used in sensitivity analysis are the to-
tal yield (loss) of nitrogen and phosphorus in runoff and with sediment, total
nitrate leached, total plant nitrogen uptake, and total denitrif ication during
1974 on watershed P2 at Watkinsvil le, Ga.

Initial estimates of parameter values used in the sensitivity analysis
were made using the procedures outlined in volume II, chapter 3. These initial
estimates are the base values; the parameters are varied about the base values
to determine model sensitivity. The hydrology and erosion input variables
(from the erosion component) are also varied about the observed values to de-
termine model sensitivity to errors in this input. Base values for the para-
meters are summarized in table 1-49.

Comparison of observed and computed nitrogen yield in runoff for 11 storm
events in 1974 produced

It = 0.039 + 0.895 N [1-215]

R2 = 0.94

140

Table 1-49. — Summary of input variables and parameters varied in nutrient sub-
model sensitivity analyses, watershed P2, Watkinsville, Ga., 1974

Variable

or Base

parameter value

Comments

Measured data

SED

Measured data

PERC

HYDONE

predictions

HYDONE

predictions

TEMP

HYDONE

predictions

AWU

HYDONE

predictions

SOLPOR

0.45

0.20

0.65

SOLN

0.20

SOLP

0.20

N03

21.0

SOILN

0.00035

SOILP

0.00018

EXKN

0.075

EXKP

0.075

16.8

-0.160

11.2

-0.146

POTM

47.0

RCN

0.80

5700

TOTAL AWU

225

TOTAL PWU

329

DMY

2.5

Daily precipitation.
Daily runoff volume.
Daily sediment yield.
Percolation.

Average soil moisture since last precipitation
event.

Average daily air temperature since last precip-
itation event.

Actual water use since last precipitation event.

Soil porosity.

Field capacity.

Organic matter in root zone.

Soluble nitrogen in surface layer.

Soluble phosphorus in surface layer.

Nitrate in root zone.

Soil nitrogen.

Soil phosphorus.

Nitrogen extraction coefficient.

Phosphorus extraction coefficient.

Nitrogen enrichment coefficient.

Nitrogen enrichment exponent.

Phosphorus enrichment coefficient.

Phosphorus enrichment exponent.

Potential mineral izable nitrogen.

Concentration of nitrogen in rainfall.

Potential grain yield.

Total actual water use, from hydrology option 1,
must be < PWU.

Total potential water use, from hydrology option
1.

Dry matter yield ratio to convert YP to total
dry matter.

141

where Ng is predicted nitrogen yield (kg/ha) in runoff and Nq is the corres-
ponding observed value. The total yields of nitrogen in runoff for the year
were 3.52 kg/ha observed and 3.58 kg/ha predicted. Therefore, the model fol-
lowed trends in the observed data and explained 94% of the variance. The cor-
responding regression equation for nitrogen yield with sediment, Ns was

fts = 0.054 + 0.810 Ns [1-216]

R2 = 0.92

with total observed yield for the year 4.3 kg/ha and computed total yield of
3.95 kg/ha. In this case, the total yields are comparable and the bias in pre-
dictions was similar to the corresponding predictions for soluble nitrogen.

The regression equation for yield of phosphorus in runoff, Pg, was

fc = 0.019 + 0.399 PQ [1-217]

R2 = 0.48.

In this case the reqression slope of 0.399 suggests a significant bias from un-
derprediction, but the relatively large intercept of 0.019 means that on the
average the predictions were not nearly as biased as the slope suggests. The
total observed yield was 0.40 kg/ha while the corresponding predicted value was
0.37 kg/ha. Although the totals are comparable, the low values for R2 and the
slope indicate that the model does not explain the trend in the data and only
explains 48% of the variance. The regression equation for phosphorus with sed-
iment was

9s = 0.024 + 0.825 Ps [1-218]

R2 = 0.91
with observed total yield of 1.505 kg/ha and computed value of 1.509 kg/ha.

From this analysis using the base values of the parameters, we conclude
that the selected parameter values produce reasonable predictions. However,
the regression results summarized above were dominated by a few large storms.

Sensitivity Analysis for Nutrient Yields in Runoff and with Sediment

The results of the surface transport sensitivity analysis are tabulated in
table 1-50, which contains information about the amount of variation and its
resulting total yield of nitrogen and phosphorus in runoff and with sediment.
For comparison, each yield column is followed by a column containing the ratio
of yield to base-value yield.

142

Table 1-50. — Nutrient submodel sensitivity analysis: chemical transport in
runoff and sediment; watershed P2, Watkinsvil le, Ga., 1974

Parameter
or

troqen

Phosphorus

Variation

variable

Base

(%)

(kg/ha)

(kq/ha)

(kn/ha)

Base

3.764

1.000

4.134

1.000

0.401

1.000

1.544

1.000

-50

74.969

19.915

4.134

1.000

0.847

2.113

1.544

1.000

-25

13.044

3.465

4.134

1.000

.425

1.061

1.544

1.000

+25

1.873

.498

4.134

1.000

.396

.989

1.544

1.000

+50

1.400

.372

4.134

1.000

.395

.985

1.544

1.000

-50

1.185

.315

4.134

1.000

.200

.499

1.544

1.000

-25

2.166

.575

4.134

1.000

.300

.749

1.544

1.000

+25

6.577

1.747

4.134

1.000

.502

1.252

1.544

1.000

+50

11.745

3.120

4.134

1.000

.603

1.505

1.544

1.000

-50

3.764

1.000

2.310

.559

.401

1.000

.854

.553

-25

SED

3.764

1.000

3.247

.785

.401

1.000

1.207

.782

+25

3.764

1.000

4.987

1.206

.401

1.000

1.868

1.210

+50

3.764

1.000

5.812

1.406

.401

1.000

2.182

1.414

-50

3.967

1.054

4.134

1.000

.401

1.000

1.544

1.000

-25

PERC

3.855

1.024

4.134

1.000

.401

1.000

1.544

1.000

+25

3.690

.980

4.134

1.000

.401

1.000

1.544

1.000

+50

3.627

.963

4.134

1.000

.401

1.000

1.544

1.000

-50

3.858

1.025

4.134

1.000

.401

1.000

1.544

1.000

-25

TEMP

3.830

1.018

4.134

1.000

.401

1.000

1.544

1.000

+25

3.607

.958

4.134

1.000

.401

1.000

1.544

1.000

+50

3.340

.887

4.134

1.000

.401

1.000

1.544

1.000

-50

4.897

1.301

4.134

1.000

.401

1.000

1.544

1.000

-25

4.147

1.102

4.134

1.000

.401

1.000

1.544

1.000

+25

3.530

.938

4.134

1.000

.401

1.000

1.544

1.000

+50

3.369

.895

4.134

1.000

.401

1.000

1.544

1.000

-25

4.981

1.323

4.134

1.000

.401

1.000

1.544

1.000

-50

AWU

6.401

1.700

4.134

1.000

.401

1.000

1.544

1.000

-75

8.148

2.165

4.134

1.000

.401

1.000

1.544

1.000

-90

9.527

2.531

4.134

1.000

.401

1.000

1.544

1.000

-50

1.420

.377

4.134

1.000

.788

1.966

1.544

1.000

-25

S0LP0R

2.226

.591

4.134

1.000

.528

1.316

1.544

1.000

+25

5.844

1.552

4.134

1.000

.331

.825

1.544

1.000

+50

8.209

2.181

4.134

1.000

.293

.730

1.544

1.000

-50

3.683

.978

4.134

1.000

.401

1.000

1.544

1.000

-25

3.738

.993

4.134

1.000

.401

1.000

1.544

1.000

+25

3.780

1.004

4.134

1.000

.401

1.000

1.544

1.000

+50

3.790

1.007

4.134

1.000

.401

1.000

1.544

1.000

143

Table 1-50. — Nutrient submodel sensitivity analysis: chemical transport in
runoff and sediment; watershed P2, Watkinsville, Ga., 1974 — continued

Parameter
or

variable

troqen

Phosphorus

Variation

Base

(%)

(kq/ha)

-50
-25
+25
+50

3.996
3.871
3.674
3.596

1.062

1.028

.976

.955

4.134
4.134
4.134
4.134

1.000
1.000
1.000
1.000

0.401
.401
.401
.401

1.000
1.000
1.000
1.000

1.544
1.544
1.544
1.544

1.000
1.000
1.000
1.000

-50
-25
+25
+50

S0LP

3.764
3.764
3.764
3.764

1.000
1.000
1.000
1.000

4.134
4.134
4.134
4.134

1.000
1.000
1.000
1.000

.204
.303
.499
.598

.509

.755

1.245

1.491

1.544
1.544
1.544
1.544

1.000
1.000
1.000
1.000

-50
-25
+25
+50

M03

3.712
3.738
3.791

3.817

.986

.993

1.007

1.014

4.134
4.134
4.134

4.134

1.000
1.000
1.000

1.000

.401
.401
.401

.401

1.000
1.000
1.000
1.000

1.544
1.544
1.544
1.544

1.000
1.000
1.000
1.000

-50
-25
+25
+50

SOL IN

3.764
3.764
3.764
3.764

1.000
1.000
1.000
1.000

2.008
3.071
5.198
6.143

.486

.743

1.257

1.486

.401
.401
.401
.401

1.000
1.000
1.000
1.000

1.544
1.544
1.544
1.544

1.000
1.000
1.000
1.000

-50
-25
+25
+50

SOILP

3.764
3.764
3.764
3.764

1.000
1.000
1.000

1.000

4.134
4.134
4.134
4.134

1.000
1.000
1.000

1.000

.401
.401
.401
.401

1.000
1.000
1.000
1.000

.772
1.201
1.887
2.316

.500

.778

1.272

1.500

-50
-25
+25
+50

EXKN

2.619
3.237
4.222
4.619

.696

.860

1.122

1.227

4.134
4.134
4.134
4.134

1.000
1.000
1.000
1.000

.401
.401
.401
.401

1.000
1.000
1.000
1.000

1.544
1.544
1.544
1.544

1.000
1.000
1.000
1.000

-50
-25
+25
+50

EXKP

3.764
3.764
3.764
3.764

1.000
1.000
1.000
1.000

4.134
4.134
4.134
4.134

1.000
1.000
1.000
1.000

.201
.301
.501
.601

.500

.751

1.250

1.499

1.544
1.544
1.544
1.544

1.000
1.000
1.000
1.000

-50
-25
+25
+50

3.764
3.764
3.764
3.764

1.000
1.000
1.000
1.000

2.067
3.101
5.168
6.202

.500

.750

1.250

1.500

.401
.401
.401
.401

1.000
1.000
1.000
1.000

1.544
1.544
1.544
1.544

1.000
1.000
1.000
1.000

-50
-25
+25

+50

3.764
3.764
3.764
3.764

1.000
1.000
1.000
1.000

2.559
3.247
5.280
6.761

.619

.785

1.277

1.635

.401
.401
.401
.401

1.000
1.000
1.000
1.000

1.544
1.544
1.544
1.544

1.000
1.000
1.000
1.000

144

Table 1-50. — Nutrient submodel sensitivity analysis: chemical transport in
runoff and sediment; watershed P2, Watkinsvil le, Ga., 1974 — continued

Parameter
or

variable

troqen

Phosphorus

Variation

Base

(%)

(kq/ha)

-50
-25
+25

+50

3.764
3.764
3.764
3.764

1.000
1.000
1.000
1.000

4.134
4.134
4.134
4.134

1.000
1.000
1.000
1.000

0.401
.401
.401
.401

1.000
1.000
1.000
1.000

0.772
1.158
1.930
2.316

0.500

.750

1.250

1.500

-50
-25
+25
+50

3.764
3.764
3.764
3.764

1.000
1.000
1.000
1.000

4.134
4.134
4.134
4.134

1.000
1.000
1.000
1.000

.401
.401
.401
.401

1.000
1.000
1.000
1.000

.994
1.237
1.931
2.421

.644

.801

1.251

1.568

-50
-25
+25
+50

P0TM

3.608
3.686
3.843
3.921

.958

.979

1.021

1.042

4.134
4.134
4.134
4.134

1.000
1.000
1.000
1.000

.401
.401
.401
.401

1.000
1.000
1.000
1.000

1.544
1.544
1.544
1.544

1.000
1.000
1.000
1.000

-50
-25
+25

+50

RCN

3.225
3.495
4.034
4.304

.857

.928

1.072

1.143

4.134
4.134
4.134
4.134

1.000
1.000
1.000
1.000

.401
.401
.401
.401

1.000
1.000
1.000
1.000

1.544
1.544
1.544
1.544

1.000
1.000
1.000
1.000

-50
-25
+25
+50

RZMAX

4.728
4.109
3.540
3.380

1.256

1.092

.940

.898

4.134
4.134
4.134
4.134

1.000
1.000
1.000
1.000

.401
.401
.401
.401

1.000
1.000
1.000
1.000

1.544
1.544
1.544
1.544

1.000
1.000
1.000
1.000

-50
-25
+25
+50

4.252
4.008
3.539
3.373

1.130

1.065

.940

.896

4.134
4.134
4.134
4.134

1.000
1.000
1.000
1.000

.401
.401
.401
.401

1.000
1.000
1.000
1.000

1.544
1.544
1.544
1.544

1.000
1.000
1.000
1.000

-25
-50
-75

-90

TOTAL
AWU

3.873
4.019
4.200
4.375

1.029
1.068
1.116
1.162

4.134
4.134
4.134
4.134

1.000
1.000
1.000
1.000

.401
.401
.401
.401

1.000
1.000
1.000
1.000

1.544
1.544
1.544
1.544

1.000
1.000
1.000
1.000

+25
+50
+75
+100

TOTAL
PWU

3.960
4.090
4.183
4.252

1.052
1.086
1.111
1.130

4.134
4.134
4.134
4.134

1.000
1.000
1.000
1.000

.401
.401
.401
.401

1.000
1.000
1.000
1.000

1.544
1.544
1.544
1.544

1.000
1.000
1.000
1.000

-50
-25
+25
+50

DMY

4.252
4.008
3.539
3.373

1.130

1.065

.940

.896

4.134
4.134
4.134
4.134

1.000
1.000
1.000
1.000

.401
.401
.401
.401

1.000
1.000
1.000
1.000

1.544
1.544
1.544
1.544

1.000
1.000
1.000
1.000

145

Table 1-51 contains a summary of the results, with the same criteria for
significance as in the hydrology and erosion sensitivity analyses.

Variations in precipitation and runoff both have significant effects on
the runoff transport of nitrogen and phosphorus, with nitrogen being much more
sensitive because of its greater downward movement. Phosphorus leaching is
relatively insignificant because of buffering action. Phosphorus in runoff is
also affected by soil porosity, soluble phosphorus, and extraction coefficient.
Nitrogen shows similar sensitivity, except to soluble nitrogen, which caused
little variation in yield. Both nutrients are sensitive to soil porosity be-
cause the nutrient accounting scheme used first calculates an initial abstrac-
tion, then allows downward movement by infiltration; the remaining nutrients
are available for runoff transport.

Constant surface soil-root zone interactions take place for nitrogen but
not for phosphorus. Therefore, nitrogen in runoff is influenced strongly by
plant water use and soil evaporation, and is influenced moderately by potential
yield, root zone depth, temperature, soil moisture, organic matter, and dry
matter yield. Nitrogen in runoff is slightly sensitive to NO3 in the root zone
potential mineralization, rainfall nitrogen concentration, percolation, field
capacity, and total plant water use.

Sediment transport of each nutrient is considered to be the product of nu-
trient content of the soil, sediment amount, and a coefficient times sediment
amount to an exponent; the sensitivity of this result is therefore easily pre-
dictable, with nutrient yields significantly responsive to changes in sediment
yield, nutrient quantities in the soil, and enrichment coefficients and expo-
nents.. It should be noted that sediment transport is a function of soil type
and is not affected by surface nutrient application or by removal during leach-
ing.

Sensitivity for Subsurface Nutrient Movement

Table 1-52 presents the results of parameter and hydrology variation in
subsurface nitrogen movement. Nitrogen uptake, denitrif ication, and nitrogen
leaching are the subsurface nutrient variables which were monitored for sensi-
tivity. These constitute the nitrogen mass balance variables in the root zone;
they are important also because of the complex interactions between the surface
and root zone nitrogen concentrations. When choosing parameter values, it is
necessary to consider the results of errors in root zone processes. Table 1-53
summarizes the significance of subsurface variable response to parameter
changes.

As would be expected, nitrogen leached is most strongly affected by field
capacity, root zone depth, and percolation. Note, however, that it is moder-
ately sensitive to 11 other parameters and slightly sensitive to 4. Uptake is

146

Table 1-51. — Results of nutrient submodel parameter and variable sensitivity
analysis, watershed P2, Watkinsville, Ga., 1974, surface transport of
chemicals!/

Parameter

Nitrogen

Phosphorus

Comments

variable

in runoff

in sediment

in runoff

in sediment

Significant

None

Significant

None

Leaching affects
soluble nutri-
ents.

Significant

None

Significant

None

Runoff affects
soluble trans-
port.

SED

None

Significant

None

Significant

PERC

Slight

None

TETF

Moderate

None

Moderate

None

AWU

Significant

None

SOLPOR

Significant

None

Significant

None

Soil porosity.

Slight

None

Moderate

None

SOLP

None

Significant

None

N03

Slight

None

SOILN

None

Significant

None

SOILP

None

Significant

EXKN

Significant

None

EXKP

None

Significant

None

Significant

None

Enrichment pa-
rameter.

None

Significant

None

Enrichment pa-
rameter.

None

Significant

Enrichment pa-
rameter.

None

Significant

Enrichment pa-
meter.

POTM

Slight

None

RCN

Slight

None

RZMAX

Moderate

None

Moderate

None

TOTAL AWU

Moderate

None

TOTAL PWU

Slight

None

DMY

Moderate

None

1/ A + 50% change in input produces a change in chemical yield of: Slight
0.01 - 10%; moderate 10 - 50%; significant > 50%.

147

Table 1-52. — Nutrient submodel sensitivity analysis: subsurface nitrogen
movement, watershed P2, Watkinsville, Ga., 1974

Nitrate
Variation Parameter leached

NL Nitrogen
Base uptake

NU
Base

Denitri- D_
fication Base

Base

(kg/ha)
33.980

1.000

(kg/ha)
124.661

1.000

(kg/ha)
37.828

1.000

-50
-25
+25

+50

23.170
29.430
36.829
39.046

.682

.866

1.084

1.149

76.877
124.661
124.661
124.661

.617
1.000
1.000
1.000

27.325
33.476
40.071
43.330

.722

.885

1.081

1.146

-50
-25
+25

+50

34.414
34.240
33.552
32.799

1.013

1.008

.987

.965

124.661
124.661
124.661
124.661

1.000
1.000
1.000
1.000

38.138
38.007
37.555
37.100

1.008

1.005

.993

.981

-50
-25
+25

+50

PERC

21.215
28.334
38.536
42.266

.624

.834

1.134

1.244

124.661
124.661
124.661
124.661

1.000
1.000
1.000
1.000

42.117
39.778
36.174
34.761

1.114

1.052

.956

.919

-50
-25
+25
+50

TEMP

33.073
33.646
32.775

30.741

.973
.990
.965
.905

124.661
124.661
124.661
112.011

1.000

.899

11.369
20.906
64.384
98.944

.301

.553

1.702

2.616

-50

-25

+25

+50_

-25"

-50

-75

-90

27.235
30.972
36.541
_38J.780_
'37~343~
41.126
45.689
49.773

.801

.911

1.075

lilil.

1~099~
1.210
1.345
1.465

124.661

.124.661

105~352"

83.107

55.402

29.218

1.000
1.000
1.000

l.ggg

"845"
.667
.444
.234

31.608

35.015

40.278

42.460_

'40~164~

42.755

45.817

48.442

.836
.926
1.065
1.123
l"062"
1.130
1.211
1.281

AWU

-50
-25
+25
+50
-50"
-25
+25
+ 50

SOLPOR

41.153
36.858
31.542
.29^382

'58~569'
43.180
27.941
23.697

1.211

1.085

.928

__.865_

1~724"

1.271

.822

.697

119.883

124.661

124~66l"

124.661

.962
1.000
1.000
1.000

ITooo"

1.000
1.000
1.000

45.686
41.054
35.230
33..001
'38~79l"
38.294
37.462
37.177

1.208

1.085

.931

__.873

1~026"

1.012

.990

.983

-50
-25
+25

+50

39.410
36.449
31.783
29.855

1.160

1.074

.935

.879

124.661
124.661
124.661
124.661

1.000
1.000
1.000
1.000

23.697
31.394
43.250
47.852

.627

.830

1.144

1.265

-50

33.934

.999

124.661

1.000

37.780

.999

-25

SOLN

33.957

.999

124.661

1.000

37.804

1.000

+25

34.002

1.001

124.661

1.000

37.852

1.001

+ 50

34.025

1.001

124.661

1.000

37.876

1.001

148

Table 1-52. — Nutrient submodel sensitivity analysis: subsurface nitrogen

transport, watershed P2, Watkinsvil le, Ga., 1974 data -- continued

Nitrate

Nitrogen

Denitri-

Variation

Parameter

leached

Base

uptake

Base

fication

Base

(%)

(kg/ha)

-50

29.119

0.857

124.661

1.000

32.839

0.868

-25

N03

31.549

.928

124.661

1.000

35.333

" .934

+25

36.410

1.072

124.661

1.000

40.323

1.066

+50

38.840

1.143

124.661

1.000

42.817

1.132

-50

34.152

1.005

124.661

1.000

37.935

1.003

-25

EXKN

34.059

1.002

124.661

1.000

37.878

1.001

+25

33.909

.998

124.661

1.000

37.784

.999

33.848

124.661

1.000

-50

28.346

.834

123.587

.991

32.479

.859

-25

P0TM

31.089

.915

124.661

1.000

35.113

.928

+25

36.870

1.085

124.661

1.000

40.544

1.072

+50

39.760

1.170

124.661

1.000

43.259

1.144

-50

33.113

.975

124.661

1.000

37.009

.978

-25

PXN

33.546

.987

124.661

1.000

37.418

.989

+25

34.413

1.013

124.661

1.000

38.238

1.011

+50

34.846

1.025

124.661

1.000

38.648

1.022

-50

45.898

1.351

124.661

1.000

31.554

.834

-25

RZMAX

39.162

1.153

124.661

1.000

35.244

.932

+25

29.945

.881

124.661

1.000

39.736

1.051

+50

26.739

.787

124.661

1.000

41.199

1.089

-50

44.670

1.315

62.331

.500

44.913

1.187

-25

39.325

1.157

93.496

.750

41.371

1.094

+25

31.117

.916

137.745

1.105

35.666

.943

+50

29.625

.872

140.806

1.130

34.266

.906

-25

TOTAL

36.087

1.062

110.632

.887

39.100

1.034

-50

AWU

38.736

1.140

93.498

.750

40.746

1.077

-75

42.801

1.260

70.126

.563

43.536

1.151

-90

46.649

1.373

47.944

.385

46.158

1.220

+25

TOTAL

38.256

1.126

99.729

.800

40.662

1.075

+50

PWU

41.107

1.210

83.107

.667

42.551

1.125

+75

43.143

1.270

71.235

.571

43.901

1.161

+100

44.670

1.315

62.331

.500

44.913

1.187

-50

44.670

1.315

62.331

.500

44.913

1.187

-25

DMY

39.325

1.157

93.496

.750

41.371

1.094

+25

31.117

.916

137.745

1.105

35.666

.943

+50

29.625

.872

140.806

1.130

24.266

.906

149

Table 1-53. — Results of nutrient submodel parameter and variable sensitivity
analyses, watershed P2, Watkinsville, Ga., 1974, subsurface transport of
nitrogen!/

Parameter

Nitrate

variable

leached

uptake

Denitrifi cation

Comments

Moderate

Significant

Moderate

Affects leaching,
so forth.

Slight

None

Slight

PERC

Significant

None

Moderate

TEW

Slight

Moderate

Significant

SCT

Moderate

None

Moderate

AWU

Moderate

Significant

Moderate

SOLPOR

Moderate

Slight

Moderate

Significant

None

Slight

Moderate

None

Significant

SOLN

Slight

None

Slight

N03

Moderate

None

Moderate

EXKN

Slight

None

Slight

Mass balance de-
termined in part
by the extraction
coefficient.

POTM

Moderate

Slight

Moderate

RCN

Slight

None

Slight

RZMAX

Significant

None

Moderate

Significant

Moderate

TOTAL AWU

Moderate

Significant

Moderate

TOTAL PWU

Moderate

Slight

DHY

Moderate

Significant

Moderate

1/ A + 50% variation in input produces a change in chemical yield of:
Slight - 0.01 - 10%; moderate 10-50%; significant > 50%.

significantly sensitive to 4 and moderately sensitive to 2 parameters, while it
shows slight or no response to changes in the other 12 parameters. Denitrifi-
cation is primarily a function of temperature and organic matter, but responds
moderately to moisture movement in the root zone.

Pesticide Submodel

This model is essentially a simplified accounting and transport model
which keeps track of pesticide concentrations on plant foliage and in the ac-
tive (1 cm) soil surface, partitions transport into the water soluble and

150

adsorbed phases, and predicts losses for up to 10 noninteractive pesticides.
That is, each pesticide is considered separately without interaction with the
plant nutrients or other pesticides.

Initial estimates of parameter values used in the sensitivity analysis
were made by R. A. Leonardi/as summarized in table 1-54. These parameter val-
ues are denoted base values, and the parameters were then varied about the base
values to determine model sensitivity. Simulations were made for a weakly ad-
sorbed pesticide (atrazine) and a highly adsorbed pesticide (paraquat). While
these chemicals are only two of many possible compounds in use, they represent
a wide range of variation in transport mechanisms and properties and thus
should be indicative of many chemicals.

Observed yields of paraquat in runoff (transport in solution phase) were
insignificant (2.) and the model predicted 0.012 g of paraquat in runoff.
Therefore, other than to say observed and computed paraquat yields in runoff
were insignificant and possibly below detection limits, no other comparisons
were made. Atrazine yields in runoff were significant and the relation between
observed yields and computed yields for 1974 and using the base values was

"Ajj = -0.03 + 0.35 AQ [1-219]

R2 = 0.63

where An is computed yield of atrazine in runoff and An is the correspond-
ing observed value. The computed values underpredicted observed yields but the
model explained 63% of the variance as indicated by R2 = 0.63. For 1974, the
observed total yield of atrazine in runoff was 8.45 g and the corresponding
predicted yield was 2.68 g. Similar predictions for years 1973 and 1975
resulted in overprediction of the total atrazine yield in runoff. However, to
be consistent with analyses for the other model components, data from 1974 were
selected for the sensitivity analysis. The 1974 observed total yield of atra-
zine with sediment was 1.18 g and the model predicted 0.22 g. The model seemed
to significantly underestimate atrazine transport with sediment.

The relation between observed and predicted yield of paraquat with sedi-
ment for individual storms was

fg = -0.38 + 0.42 Ps [1-220]

R2 = 0.88

where Ps is the computed yield of paraquat with sediment and Ps is the cor-
responding observed yield. Although the model explained 88% of the variance,
the total observed paraquat yield with sediment was 102.2 g and the model pre-
dicted a total yield of 40.0 g. Again, there is a significant underprediction
for 1974 but the model does explain the trend in the data. The yield of para-
quat with sediment is a function of the calculated enrichment factor. This

4/ Soil scientist, USDA-SEA-AR, Athens, Ga., personal communication

151

Table 1-54. — Summary of input variables and parameters varied in pesticide
submodel sensitivity analyses, watershed P2, Watkinsville, Ga., 1974

Variable

or
parameter

Base
value

Measured
data.

SED

Measured
data.

ENRICH

Predicted by
erosion model

SOLPOR

0.45

DEPINC

1.00

EFFINC

1.00

SOLFRC

1.00

S0LH20

33.0 Ai/
5xl05 P

EXTRCT

0.10

DECAY

0.14 A

0.007 P

4.00 A
106 P

APRATE

3.40 A

2.049 P

Comments

Daily precipitation.

Daily runoff volume.

Daily sediment yield.

Sediment enrichment ratio.

Soil porosity.
Depth of incorporation.
Efficiency of incorporation.
Fraction applied to soil.
Water solubility, ppm.

Extraction ratio.
Decay constant.

Rate of chemical application,

1/ A refers to atrazine and P refers to paraquat.

enrichment factor is calculated based on organic matter content and selective
deposition of larger size sediment fractions. The relative accuracy of these
calculations is unknown.

Analysis of model predictions using the base values of the parameters sug-
gests that predictions of weakly adsorbed atrazine yields with runoff and of
highly adsorbed paraquat yields with sediment explain trends in the observed
data. In both cases, the model underpredicts total yields. However, the model
did not produce reasonable estimates of atrazine yields with sediment or para-
quat yields with runoff. Nevertheless, the results represented by equations
[1-219] and [1-220] are precise enough to enable analyses to determine model
sensitivity.

152

Sensitivity Analysis for Pesticide Yields in Runoff and With Sediment

Total yield of atrazine and paraquat in runoff and with sediment were cal-
culated for the 1974 data. As each parameter (or input variable) was varied a-
bout the base value, the total yields were compared to the totals obtained
using the base values as a measure of the sensitivity. Column 2 of table 1-55
shows which parameters were varied and by how much. Columns 3, 5, 7, and 9
show the predicted yields and columns 4, 6, 8, and 10 show the predicted yields
divided by the corresponding yields using the base values of all parameters.

As expected, yield of atrazine in runoff was sensitive to changes in rain-
fall and runoff and not sensitive to changes in sediment yield or enrichment
(see table 1-55). Notice that each input variable was varied independently, so
that a decrease in rainfall increased the pesticide yields because the runoff
remained unchanged and thus proportionally more rainfall became runoff. On the
other hand, the yield of paraquat with sediment was sensitive to sediment yield
and enrichment. The sensitivity of pesticide yield with sediment to rainfall
varies with the distribution coefficient KD. For atrazine, the base value is
KD = 4.0 and for paraquat KD = 10°; atrazine yield with sediment was sensi-
tive to total rainfall while paraquat was not.

^Jery significant parameters are: EFFINC, the efficiency of incorporation
of the pesticide into the soil; DEPINC, depth of pesticide incorporation into
the soil; and SOLFRC, fraction of the pesticide applied directly to the soil.
This is fortunate in the sense that these parameters can be accurately deter-
mined but unfortunate in the sense that detailed knowledge of pesticide appli-
cation and cultivation techniques must be known and these vary with soil condi-
tions.

The distribution coefficient, KD, is a parameter representing the distri-
bution of pesticides between solution and soil phases. A low value of KD re-
presents relatively more of the pesticide in solution while high values of KD
represent the opposite. Various values of KD should be compared in terms of
their order of magnitude rather than small differences, especially for large
values. For small values of KD < 10, small differences may be significant.
This can be seen by comparing the relative sensitivity of atrazine and paraquat
yields for KD variations, in table 1-55.

Sensitivity of the pesticide submodel is summarized in table 1-56. As
described at the bottom of the table, the most sensitive input variables and
parameters are indicated as "significant" in table 1-56. For the input vari-
ables and parameters listed as significant, errors in estimating the input or
parameter values are magnified by the model. With this criterion, 10 of the 15
parameters and input variables listed in table 1-56 are significant with res-
pect to error magnification. Also, if KD is varied by orders of magnitude (as
it must be to represent a broad spectrum of pesticides) then it, too, is sensi-
tive. Compared to the hydrologic and erosion/sediment yield components, this
is a relatively large number of "significant" parameters. This has the posi-
tive aspect that if a model is to be useful in reflecting management practices,
it should be sensitive to parameters representing the practices. However, when
more parameters are required to a higher degree of precision, then more effort
and information are required to provide input to the model.

153

Table 1-55. — Pesticide submodel sensitivity analyses, watershed P2, Watkins-

ville, Ga., 1974

Atrazine

Paraquat

with

runoff

sediment

runoff

sediment

Paramete

sr Aq

AQ/

AS/

PQ/

PS/

Variation

base

(1)

(2)

(3)

(4)

(5)

(6)

(7)

(8)

(9)

(10)

(%)

(£)

(1)

(!)

Base

2.678

1.000

0.215

1.000

0.012

1.000

40.041

1.000

-50

4.718

1.762

.278

1.293

.012

1.000

40.042

1.000

-25

3.574

1.335

.247

1.149

.012

1.000

40.042

1.000

+25

2.030

.758

.188

.874

.012

1.000

40.041

1.000

+50

1.577

.589

.166

.772

.012

1.000

40.041

1.000

-50

1.295

.484

.213

.911

.006

.500

40.041

1.000

-25

1.975

.737

.214

.995

.009

.750

40.041

1.000

+25

3.405

1.271

.216

1.005

.014

1.167

40.041

1.000

+50

4.153

1.551

.216

1.005

.017

1.417

40.041

1.000

-50

2.678

1.000

.107

.498

.012

1.000

20.086

.502

-25

SED

2.678

1.000

.163

.758

.012

1.000

30.084

.751

+25

2.678

1.000

.270

1.256

.012

1.000

49.983

1.248

+50

2.678

1.000

.322

1.498

.012

1.000

59.866

1.495

-50

2.678

1.000

.107

.498

.012

1.000

20.086

.502

-25

ENRICH

2.678

1.000

.161

.749

.012

1.000

30.092

.752

+25

2.678

1.000

.268

1.247

.012

1.000

49.969

1.248

+50

2.678

1.000

.322

1.498

.012

1.000

59.865

1.495

-50

2.845

1.062

.215

1.000

.012

1.000

40.041

1.000

-25

SOLPOR

2.771

1.035

.215

1.000

.012

1.000

40.041

1.000

+25

2.527

.944

.214

.995

.012

1.000

40.041

1.000

+50

2.296

.857

.212

.986

.012

1.000

40.041

1.000

-50

5.356

2.000

.429

1.995

.023

1.917

80.083

2.000

-25

DEPINC

3.571

1.333

.286

1.330

.015

1.250

53.389

1.333

+25

2.142

.800

.172

.800

.009

.750

32.033

.800

+50

1.785

.667

.143

.665

.008

.667

26.694

.667

-100

.000

-75

EFFINC,

.670

.250

.054

.250

.003

.250

10.010

.250

-50

SOLFRC

1.339

.500

.107

.500

.006

.500

20.021

.500

-25

2.009

.750

.161

.750

.009

.750

30.031

.750

154

Table 1-55,

'esticide submodel sensitivity analyses, watershed P2, Watkins-
ville, Ga., 1974 — continued

Atrazine

Paraquat

with

runoff

sediment

runoff

sediment

Parameter Aq

AQ/

As/

PQ/

ps/

Variation

base

(1)

(2)

(3)

(4)

(5)

(6)

(7)

(8)

(9)

(10)

(%)

(S.)

(jsO

(£)

xlO-2

5.280

1.972

.915

4.260

.012

1.000

40.041

1.000

xlO"1

S0LH20

2.678

1.000

.215

1.000

.012

1.000

40.041

1.000

xlO

2.678

1.000

.215

1.000

.012

1.000

40.041

1.000

XlO2

2.678

1.000

.215

1.000

.012

1.000

40.041

1.000

-50

1.563

.584

.125

.582

.012

1.000

40.041

1.000

-25

EXTRCT

2.163

.808

.174

.810

.012

1.000

40.041

1.000

+25

3.124

1.167

.251

1.169

.012

1.000

40.041

1.000

+50

3.514

1.312

.282

1.313

.012

1.000

40.041

1.000

-50

12.232

4.568

.444

2.067

.015

1.250

50.380

1.258

-25

DECAY

5.143

1.920

.290

1.350

.013

1.083

44.745

1.117

+25

1.550

.579

.166

.773

.010

.833

35.667

.891

+ 50

.998

.373

.131

.610

.009

.750

32.033

.800

100

.983

.367

.266

1.237

110.873

9.2xl03

2.346

.059

ioi

KDl/

2.964

1.107

.457

2.128

197.277

1.6xl04

7.615

.190

102

.744

.278

.959

4.465

92.516

7.7x103

32.604

.814

103

.084

.031

1.061

4.939

11.258

9.4xl02

39.203

.979

104

.009

.003

1.072

4.991

1.149

95.750

39.957

.998

-50

1.339

.500

.107

.498

.006

.500

20.030

.500

-25

APRATE

2.009

.750

.161

.750

.009

.750

30.036

.750

+25

3.348

1.250

.269

1.252

.014

1.167

50.047

1.250

+50

4.017

1.500

.322

1.499

.017

1.417

60.072

1.500

1/ Figures for KD do not represent relative changes but actual KD values
used from 10u to 104.

The pesticide submodel represents extreme simplifications of complex nat-
ural processes. Although the number of "significant" parameters is relatively
large, the prediction results summarized by equations [1-219] and [1-220] indi-
cate that additional simplifications would result in an oversimplified model
which would not represent the known processes as well as this model does.
Therefore, it seems likely that a model of at least the complexity used herein
will be required to predict pesticide losses with improved accuracy.

155

Table 1-56. — Variables and parameters varied in pesticide sensitivity analysis,
watershed P2, Watkinsville, Ga., 1974

Parameter

or
variable

Atrazine
in

runoff

Atrazine

with
sediment

Paraquat
in

runoff

Paraquat

with
sediment

Comments

P
Q

Significant!/
Significant

Significant
Slight

None
Significant

None
None

Sensitivity de-
pends upon the
value of KD for
hydrologic in-
puts.

SED

None

Significant

None

Significant

ENRICH

None

Significant

None

Significant

Measure of silt-
clay enrichment
important for
adsorbed chemi-
cals.

SOLPOR

Moderate

None

More significant
for soluble
chemicals.

DEPINC

Significant

Highly depen-
dent on soil
conditions.

EFFINC

Significant

Highly depen-
dend on soil
conditions.

SOLFRC

Significant

Highly depen-
dent on soil
conditions.

S0LH20

None

EXTRCT

Significant

None

DECAY
KD

Significant
11

Moderate
2/

Moderate
11

Distribution co-
efficient most
sensitive for
small values of
KD.

APRATE

Significant

All yields sens-
itive to appli-
cation rate.

1/ A + 50% change in parameter value produces a change in chemical yield
of: Slight < 10%; Moderate 10-50%; Significant > 50%.

2/ Parameter varied by orders of magnitude; sensitivity relationship non-
linear but significant for large changes.

156

SUMMARY

Sensitivity analyses were conducted for the field scale model using data
from watershed P2 at Watkinsville, Ga. Sensitivity to input data and parameter
values was determined by considering the hydrology, erosion/sediment yield, nu-
trient, and pesticides submodels separately. Because of the complexity, inter-
actions, and number of simulation runs required, no attempt was made to deter-
mine sensitivity of the entire model involving the simultaneous operation of
its components. However, insight for sensitivity of the entire model can be
gained by considering sensitivity of its components and their linkage in the
field scale model .

Output from the hydrology component provides input to the erosion/sediment
yield component. Both these components in turn provide input to the nutrient
and pesticide components. Observed rainfall data were used in determining
sensitivity of the hydrologic components. Observed rainfall and runoff data
were used to determine sensitivity of the erosion/sediment transport component.
Observed rainfall, runoff, sediment, and climatic data were used to determine
sensitivity of the nutrient and pesticide submodels. By using observed data
where possible, we sought to minimize compound errors and interactions due to
errors in predictions from the hydrology and erosion/sediment yield components.
Also, it is important to stress that model simulations represent predictions
using base values of the parameters and that no attempts were made to optimize
or calibrate model components using the observed data. Observed data were used
for comparison and to evaluate model predictions.

The quality or accuracy of the model predictions made using base values of
the parameters represent the type and magnitude of errors which might be expec-
ted in applying the model to predict runoff, sediment yield, nutrient losses,
and pesticide losses on a complex agricultural watershed. Moreover, conclu-
sions regarding model sensitivity refer to the results for a specific watershed
with initial parameter estimates derived using procedures outlined in the
volume II, User Manual. Somewhat different results could result from applica-
tion of the model under different conditions.

A qualitative assessment of the significance of and sensitivity to input
variables and parameters for each component or submodel was made using the cri-
terion that the sensitivity to a particular parameter is "significant" if er-
rors in that parameter result in errors in the submodel output as large or lar-
ger than the parameter errors. In this case, the model was said to magnify the
errors. Sensitivity assessments for the hydrology component are listed in
tables 1-37 and 1-40, and for the erosion/sediment yield component in table
1-44. Similar assessments for the nutrient submodel are given in tables 1-51
and 153, and for the pesticide submodel in table 1-56.

REFERENCES

(1) Laflen, J. M., H. P. Johnson, and R. 0. Hartwig.

1978. Sedimentation modeling of impoundment terraces. Transactions of
the American Society of Agricultural Engineers 21(6): 1131-1135 .

157

(2) Smith, C. N., R. A. Leonard, G. W. Langdale, and G. W. Bailey.

1978. Transport of agricultural chemicals from small upland Piedmont
watersheds. Environmental Protection Agency, EPA-600/3-78-056. U.S.
Government Printing Office, Washington, D.C. 364 pp.

(3) U.S. Department of Agriculture, Soil Conservation Service.

1972. National Engineering Handbook, Sec. 4, Hydrology. 548 pp.

158

CREAMS

A Field Scale Model for

Chemicals, Runoff, and Erosion From

Agricultural Management Systems

VOLUME II. USER MANUAL

CONTENTS

Chapter Page

Introduction- -------------------------- 161

— W. G. Knisel and J. D. Nowlin

1 Hydrology ---------------------------- 165

—J. R. Williams, R. E. Smith, J. D. Nowlin, and A. D. Nicks

2 A model to estimate sediment yield from field-sized areas: - - - 193

selection of parameter values

--G. R. Foster, L. J. Lane, and J. D. Nowlin

3 Nutrient submodel 282

— M. H. Frere and J. D. Nowlin

4 Pesticide submodel- ----------------------- 304

--R. A. Leonard and J. D. Nowlin

5 Example applications for typical field situations -------- 330

— G. R. Foster, M. H. Frere, W. G. Knisel, R. A. Leonard,
A. D. Nicks, J. D. Nowlin, R. E. Smith, and J. R. Williams

160

VOLUME II. USER MANUAL

W. G. Knisel and J. D. Nowlin^

INTRODUCTION

CREAMS, volume II, is structured as an independent publication written for

(1) the technical user to develop input and parameter information and (2) the

computer programer to set up data files for running the model. Reference ta-
bles and figures in this section are for the user's convenience.

CREAMS is structured as three separate components: (1) hydrology, options
one and two; (2) erosion/sedimentation; and (3) chemistry, plant nutrients, and
pesticides. Although the influences of agricultural management systems on run-
off, erosion and sediment transport, and chemical transport are complex and in-
teractive, the user may wish to isolate these influences for specific compon-
ents. A practice may have only a secondary influence on the hydrologic perfor-
mance of a watershed, for example, while having a major influence on sediment
yield or chemical transport. Rerunning the hydrologic component for every al-
ternate practice considered for erosion control therefore is unnecessary. Re-
running the hydrology and erosion components to evaluate the effects of split
applications or alternate pesticide applications also is unnecessary. Indepen-
dent operation enables the user to consider management options and make compar-
isons while operating at a relatively low cost. The user also may want to run
only the hydrology and erosion components rather than the chemistry component.

Running the model as three separate components places some restrictions on
the user, who must record the files generated by a component to pass the cor-
rect file to the next component. Although this problem is not severe, the user
should be aware of the many files that can be generated rather quickly for a
specific problem.

Figure 1 1 - 1 is a generalized chart of program flow that shows input and
output files and pass files from one component to the next and the sequence of
operations. As shown on the left side of figure II-l, the users also can
supply the hydrology or sediment yield data, or both, if they so desire. If a
historic record of runoff and sediment yield data is available for some loca-
tion near the farm of interest, for example, the users might want to use obser-
ved runoff and sediment yield data and run the chemistry component rather than
generate runoff and sediment yield with the model.

1/ Hydraulic engineer, USDA-SEA-AR, Southwest Range! and Watershed Research
Center, Tucson, Ariz., and computer programer, Agricultural Engineering Depart-
ment, Purdue University, West Lafayette, Ind., respectively.

161

PRECIPITATION
DATA (FILE 4)

HYDROLOGY

PARAMETERS DATA

(FILE 5)

' '

HYDROLOGY
PROGRAM

HYDROLOGY
OUTPUT REPORT
(FILE 6)

DGY PASS A
FILE 7) )

USER SUPPLIED
HYDROLOGY DATA

EROSION/SEDIMENT

YIELD

PARAMETERS DATA

(FILE 5)

EROSION/SEDIMENT
YIELD PROGRAM

EROSION/SEDIMENT

YIELD OUTPUT

REPORT (FILE 6)

HYDROLOGY

SEDIMENT YIELD

PASS DATA

(FILE 7)

USER SUPPLIED

HYDROLOGY

SEDIMENT YIELD

DATA

CHEMICALS

PARAMETER DATA

(FILE 5)

CHEMICALS
PROGRAM

CHEMICALS

OUTPUT REPORT

(FILE 6)

Figure II-l. — Schematic representation of flow of CREAMS programs

162

Each component of CREAMS and its input and output are treated separately
in subsequent chapters. Sample data and parameter values are given in the res-
pective chapters with schematic representations of card decks showing card
order, variable name, and format on each schematic card image. Tables are in-
cluded to show parameters, parameter definition, source of information, and
relative accuracy of the estimation of parameters. The output information
shows specified options to print the desired information. Samples of output
are shown, and each output element is defined.

Chapter 5 gives examples of typical computer runs for three specific field
situations: (1) Georgia Piedmont, (2) Mississippi Delta, and (3) western Ten-
nessee. These situations provide a wide range of soils, topography, management
practices, and climate. Specific parameter values are shown as well as some
typical output. This information and description will help the user to under-
stand and follow through the model operations.

COMPUTER REQUIREMENTS

The computer programs are written in standard FORTRAN IV. They were writ-
ten and tested initially at Purdue University on a CDC-6500. They also were
tested on a DEC-10 at the University of Arizona in Tucson and the IBM 370/158
at the U.S. Department of Agriculture, Washington Computer Center in Washing-
ton, D.C. Several other locations are testing experimental versions of indivi-
dual programs in the model. These programs can adapt to hardware configura-
tions with only a few minor modifications.

All three programs will compile, load, and execute in less than 55K words
of memory on the CDC-6500. This figure varies according to the installation,
but the programs can be run on a relatively small machine. Significant digits
are hardware dependent but do not affect the numeric variable formats. Since
alphanumeric variable formats can be affected, so they were kept small (A4) to
facilitate the use of smaller computers.

The computer programs require the use of secondary input/output devices
and two input files that can be accessed independently. The first and second
programs in the series also require an extra output device to handle the pass
files generated for the next program. Table 1 1 — 1 estimates storage require-
ments and compilation, load, and execution time.

163

CM I

CSJ|

t— I <— I OO

cr> o o

.— 1 uo <x>

O VJO <T>

IT) O

O *d- CTi

UO «* CO

UD CO OJ

ro cr>

C\J CO vo

■3- «-H rH

UD *d" CO

un un

i— I CVJ

<T3 00

O LT)

< — 1

_) <—<

*•>. o

^-\ O

*>*. O

*-». o

o o

-MOO

o o

r-. un O

•r- r^ UO O

NlfiO

r^. un

O) CO UD i—l

UJ 1 1 1

■^s s: <_> <_>

co ud <— i

co ko

i- 2: o c_>

s: o o

v^.

2: o

CLCQ Q UJ

-*: CO O UJ

mou

-i

0Q O

E •— ' C_> Q

c •— ' o o

-M

-h O O

i— < <_>

•'—

o -o

+_>

rt3

4->

ai fd
o

sz c
a: a:

rH|cM|

164

Chapter 1. HYDROLOGY

1/
J. R. Williams, R. E. Smith, J. D. Nowlin, and A. D. Nicks

INPUT DATA FILES

The information in this chapter will help the user assemble the precipita-
tion data files, temperature and radiation files, and hydrology parameter
files. As described in CREAMS, volume I, chapter 2 (Hydrology), two options
are available to the user. Option 1 uses daily rainfall and requires 37 cards
per year of precipitation data with 10 daily values per card, as described in
table 1 1-2 - A schematic deck arrangement is shown in figure 1 1-2 - The format
in figure 1 1-2 and table 1 1-2 (that is, 10X, 10F5.0) represents a read format
for data, only. Sample data in table 1 1-2 show the year, 74, in columns 4 and
5 and sequential card numbers within the year in columns 79 and 80. These are
for identification of data only and not for use in the programs. Daily rain-
fall data are available from the climatological data of the National Weather
Service and from several USDA-SEA-AR research locations.

Formats for breakpoint rainfall data are given in table 1 1-3, and a sample
deck arrangement is shown in figure II-3. Breakpoint rainfall data are avail-
able upon request from USDA-SEA-AR for several locations in the United States.
As described in CREAMS, volume I, chapter 2, hourly rainfall data can be used
as input for hydrology model option 2. Hourly data are available from the
National Weather Service for many locations in the United States for the period
since 1948. Hourly rainfall data would be input in the same format as that
used for the breakpoint data. The hourly time entries would be in clock hour.

Users of the CREAMS model interested in running hydrology model option 2
who do not have breakpoint rainfall data should contact USDA-SEA-AR locations
in their respective States for availability of breakpoint rainfall data. To
evaluate management practices, rainfall data can be transferred some distance
within climatic regions. In mountainous regions, orographic influences must be
recognized. Since evaluation of management systems is relative for a given
climatic record, the period of record is unimportant and records from 1930 to
1940 would be just as appropriate as data from 1968 to 1978. Since representa-
tiveness is important, the data should include years with above normal, near
normal, and below normal annual rainfall so that results and interpretations
are not biased unduly. Selection of record period may be critical for areas
west of the 100th meridian. If the user has a suitable method of generating
synthetic data, either daily or hourly, such data would be entirely satisfac-

1/ Hydraulic engineer, USDA-SEA-AR, Grassland-Soil and Water Research
Laboratory, Temple, Tex.; hydraulic engineer, USDA-SEA-AR, Fort Collins, Colo.;
computer programer, Agricultural Engineering Department, Purdue University,
West Lafayette, Ind.; and hydraulic engineer, USDA-SEA-AR, Southern Great
Plains Watershed Research Center, Chickasha, Okla.

165

Table 1 1-2 - — Daily precipitation data input files

Precipitation Data File

For Daily Rainfall model (option 1)

Card 1-37. R(l-365)

R() Daily rainfall (in) , e.g. 2.07

The rainfall data are on a separate file from the parameters. They're
read in, 10 values per card, 37 cards per year, and are repeated for each year
of simulation. The year along the left margin and the seguence number along
the right margin are only to aid the user in putting together the data. The
program doesn't read them in. The following sample is only a partial years
data.

Format(10X,10F5.2)

2.07

.16

.37

.17

0 .34

.08

0 0

.87

.24

0 0

.10

.26

0 1.70

.20

.65

.90 0

.16

REMAINDER OF THE

DAILY PRECIPITATION DATA

(10 DAYS/CARD. 3? CARDS/YEAR)

FDRNAT<10X>10F5.0)

Rll R12 R13 R14 R15 R16 R17 R13 R19 P20

Rl RE R3 R4 R5 R6

R8 R9 R10

Figure II-2. — Sample deck arrangement for daily rainfall input for hydrology

model option 1.

tory, and the generation scheme could be modified for output compatible with
the input data formats described in tables II-2 and 1 1-3.

Precipitation data are available to the user from different sources. Dai-
ly and hourly rainfall data are published by the National Weather Service (NWS)
(_7, 9). These data also are available on maqnetic tape and can be purchased
from the National Weather Data Center, Asheville, N. C.

Breakpoint rainfall data, required by hydrology model option 2, are avail-
able for several SEA-AR research watershed locations across the United States.

166

Table 1 1-3.— Breakpoi nt rainfall input data files

por Breakpoint Rainfall Model (option 2)

Card 1. JYR, JDAY, NP, MIDNI, PRE(JDAY)

JYR Year of event (last 2 digits), e.g. 74

JDAY Day of event (Julian day), e.g. 001

NP Number of breakpoints in the event, e.g. 6

MIDNI 0 if event takes place during only one day
1 if event overlaps into two days

PRE() Total rainfall for event (in), e.g. 2.07

Card 2. BP(l-NP) , T(l-NP)

BP() Accumulated rainfall at time T() (in), e.g. 1.96

T() Time of measurement (min. from midnight), e.g. 38.0

The rainfall data are on a separate file from the parameters. Card 2 is
repeated for each breakpoint (NP, card 1) during the event. A card 1 and a
series of card 2's are repeated for each event during the simulation. A par-
tial sample, two events, follows.

Format(4l8,F8.0)

Format (2F8.0)

0 2.0700

0.0

1.0

.9600

38.0

.0300

47.0

.0400

54.0

.0500

154.0

.0700

180.0

O.U

1213.0

.0500

1215.0

.0800

1217.0

.0900

1223.0

.1600

1230.0

U .1600

167

< FLAGS THE END DF FILE)

REMAINDER DF THE

EREAK-POIHT PRECIPITATION DATA

(NP+1 CARDS/EVENT)

FDRMAT(4I3)F3.0>
FDRMAT<F3.0.I3>

EP<NP> TCNP)

Figure 1 1 -3 . — Sample deck arrangement for breakpoint rainfall input for
hydrology model option 2.

A user might want to transfer these data for application within climatic
regions, these data files are available in standard format on magnetic tape or
cards at the Water Data Laboratory, USDA-SEA, Beltsville Agricultural Research
Center, Beltsville, Md. 20705. The standard format of these data is different
from the model input reguirement and must be processed to conform to nradel in-
put.

Several data sets were assembled for USDA-SEA research watersheds to test
the CREAMS hydrology component. These data sets are available from the Water
Data Laboratory and include runoff, air temperature, and solar radiation data,
(table 1 1-4) . Data on soils, land use, and manaaement have been published for
these watersheds (4) .

Hydrology Options

The hydrology submodel operates on a given rainfall data seguence plus a
record of average monthly radiation and temperature, with information on crops,
soil profile, and field shape to generate a seguence of information on runoff,
evaporation, and seepage. This output information is used by the erosion, pes-
ticide, and nutrient models in simulating chemical transport.

The hydrology model is desiqned to use physically related or easily estim-
able parameters as much as possible. It does not depend on extensive detail
for soil or field topography. Plant growth patterns for crops grown are speci-
fied for a normal situation but are modified within the model for extreme
stress (drought) conditions.

The simplifications used are dictated largely by data limitations rather
than ignorance of the interrelations of the physical processes involved. Major
limitations are:

(1) If only daily rainfall records are available, runoff is estimat-
ed by the SCS curve number (CN) procedure. Peak runoff rate

168

OCOCOtCCMCMtvN

^f *d" VO VO

r~- r — i — . r^
vo vo 10 vo

en ID ID VO <D

ID VO VD VO

VO VD VO ID LO I

VO VO VD ID

&\ O O CM O O O l

!E!E<E!E!E£!Efe!E£8£!iEtE£

CM CM

VD ID ID VD
VD <D VD VD

ID ID
CM CM

oo ro

VO <D

ID ID
ID ID

6674
6674
6676
6676

<D id

CO 00
ID VO

ID <D

VD VO
ID VO

s: si

vn r^
rv. vc

CO CO

VO vn

CD6774
CD6774
CD6774
CD6774
MT6676
PT7377
PT7277
PT7277
PT7277
MT6972
MT6975
MT7275
MT7375

ICUlHOlOONONNNMniOlOvOVO
^NNVONMsNvOVOlOirUlNMsr^

Noococoa>oiooiooooocomco»co

O000O0^-POP0«d'0O»a-«a-"3-O-)rovDVDVDVD

VD VD LD I
VO VD (

£££££££!

CM ID LD ID

£££

VD VO VD VD

££££

.(_> j_> 4_> S_ TO O

S. S. S- I. £

322imul-

v- V-
O O

+-> +->

c c
o o

t_> CJ

V- s_

> Q. •

1 +-> -l->
J 1/0 o :

vn +j cr. c vD

C I- +-> CD • i— C C CT CD C —

-l-s-foaiOi:xr>ro'D33-'-oooo

)l-wn:au3LuvDci:Q:viioCLOoi/i"

<_> O Q

U O i

i- r- cccccncr, cc cc5S

CT) C7> D"i CT) C7> CT) »r- •*-OOOOCC"0"OCV_> -i— •»— oo

CCC<UCvDCC4->-U+J>>4J+J4->.t->,'-,'-,'-,'-QJ ID « L. b

.*-*.*c.*c.*.i:ccccc:</>t/)v/i</i.i->4->>>(DN'— r— m- it-
iflmi/)aiui<UwtoiT3iT3'Dx:-c:3Zi33wvnoo>-t-'<_)C_Jcc
3H3 0J3 3Lt.l.t)tJOOOO«f|01.1.«'i-UUa)(l)

■ — o- a — i _i sr s: cc cc

oooon w:

^ ro o c +->

t- O -* O S- i— i—

S i— 0) >» * *- -i-

•>-•-,— c 5 o o

(O ra o CD O CD CD

^d. vy uuu

CMCMCMCMCMCMCMCMOOO^xd-VDVOVOVO-— I I

HrHHiH^lrHHrfCOCOCOCOCOCOCOCOCOM

nrt«a-oooocr>co

•Hfl-rHcocococorom

O.OI/1IIIIDD

oocn.-ir^voaiiD>oicr>.— i nocoMrxj-i

VOVOVOCMVOLOr^CTlO'lCMLO • • • CT> CX) VT» '
VO O O • • •

.-hO.— ii— ii-nr-^r-ir-4i— irocM-— i ^i- cm cm m co >

*J- CTi O CM ID O '

OcvinnHH

r-(<jirini

3 3 3 3 >-

v/) ro cm 3 3

O or cc 3

IIIIIIII
OOOOOOOOi^^^i

O O O i<i ^
CCCCCCCC OOXXX

OOOOOOOOVTJCDCD |_|— (_

+->-|J+->-l->+->4->+J+JCDCD<DCDCD

o vj u o o o (_> o .* -* .* •!-■!- r— i— r- i

OOOOOOOOOOOl-S-CDCDCD

N3^iii;^ ro CDCD

«C OOOOC CD CT> i— r—

■D O O 3 >— r—

OO GO CD ' — (O fO <0 (T3 +J | — .COO «=£•!—•(—

— c (urrr r cz qj < cs>>

cr> cr>
c c "O "O +-> it-
l/> CD .

-CS-l-(-.e.ci/>vv>i/>i/>+J.t->OO.Qi-

i/ii/)i/imi/i(/iv/il/»CDCDCD+->-4->CDCDCDCD

000000004Ex:x:33,'->'-,'-,'-«s«8xxoo^:x:jcjc+J->-<oi03ET-ion3

uuuuouuuuuuc3(3Q;o:a:Q:iiooi-suoouii-HJ3i:<i-33

o re :n •
ro <o +j >-,.c 3 i-t c c e

^^ l/) QJ IOi — <T3 0*'"'r*

oo c ail''- w+j^ji

T3 3 ■«" C CDM-j-J+J

wouia)H^oinkONnwio<j-u)iD9Mn'*i«inHfo^c\j^woooooooo

OOOOt— I.— ICMCMOO'— lOOi— IHr- ICMOCM^H^HOOOO<r>"d-«S-000000000
OOOOOOOOOOOOOOOOOOOOO'-HOOOOOO
^lDVDCMCMCMCM>a-«3-CMCMrovDCriCrvCrvCTlO0
CO OO ^^■^■^t^'^'VOVOVOUJVOlDVOVON

VD VD VD VD ID VD VD VO ^" ^"

CM CM CM CM CM CM CM CM r- —

169

and rainfall erosive energy index (EI) are predicted with re-
gression equations based on runoff volume and watershed charac-
teristics.

(2) Average daily values of net radiation and temperature are used
for all years of a simulation. Since evapotranspiration (ET)
depends strongly upon radiation and temperature, the standard
deviation of ET is underestimated. The average radiation and
temperature values give good estimates of long-term averaae ET,
however.

(3) The soil profile is assumed to be constant in hydraulic proper-
ties throughout the growing season and constant (but different)
in properties throughout the fallow period. This necessarily
ignores the specific changes due to cultivations, rain splash
crusting, and other time variations in soil properties. Model
simulation methods to account for the effects of cultivation on
infiltration and other properties of soil water movement await
the results of current research.

(4) Soil water is assumed to move downward as a simple linear
threshold model so that the elements of soil storage transfer
water downward by gravity only when field capacity is exceeded.
This simplification is necessary since the nonlinear differen-
tial equations for unsaturated water flow require input informa-
tion and computational complexity far beyond the needs and re-
sources of this management model.

Operation

Figures 1 1 -4 and 1 1-5, flow charts for the hydrology simulation models,
use the daily and breakpoint-infiltration options, respectively. These models
operate on a time step of one day and use an alternate runoff model only on
those days when rainfall occurs. The SCS-CN method or the infiltration method
may be used for runoff simulation, depending on available rainfall data. Some
hydrologic parameters that the user must specify will be different for these
two models. Others will be common for either option. The hydrology model
takes parameter input data from the parameter file and operates sequentially as
precipitation information is read from the precipitation input file.

Input parameters necessary are listed in table 1 1 - 5 . This table refers to
equations in the documentation section, volume I, chapter 2, where appropriate,
and gives definitions and sources for parameter values.

Computer card format with variable names and definitions is shown in table
1 1-6. The hydrology program reads from one file containing parameter values
and another containing the actual precipitation data.

All numeric input formats consist of 8 column fields unless specifically
stated otherwise in the card description. Integers are read with 18 formats,
and real numbers are read with F8.0 formats. Integers must be right justified
in columns 1 to 8, 9 to 16, 17 to 24, ... 73 to 80. Real numbers must be

170

f START J

ATTACH RAINFALL
RUNOFF DATA

ATTACH WATERSHED
PARAMETER DATA

INITIALIZE ALL
PARAMETERS

CALCULATE DAILY

AVERAGE TEMP (°C) AND

RADIATION, AND LEAF

AREA INDEX

(

READ ONE YEARS VALUES
OF DAILY RAINFALL

1 = 1 + 1

CALCULATE ANY SNOWMELT,
OR ADO TO SNOWPACK

OVER
DAYS

| COMPUTE RUNOFF

LOOP
369

COMPUTE ET, SOIL WATER
MOVEMENT AND SEEPAGE

| ESTIMATE PEAK FLOW |

NM>wK

Figure 1 1-4. — Generalized flow chart for HYDONE (hydrology option 1)

171

( START )

ATTACH RAINFALL
RUNOFF OATA

CALCULATE OVERALL
STATISTICS

ATTACH WATERSHED
PARAMETER OATA

INITIALIZE ALL
PARAMETERS

CALCULATE DAILY

AVERAGE TEMP. (°C) ANO

RADIATION, AND LEAF

AREA INDEX

READ ONE OAILY RECORD

CALCULATE NO. DAYS
SINCE LAST RAIN (IDO)

i 1*1

YES

^YEAR ENCOUNTERED^"

CALCULATE ANNUAL

STATISTICS, WATER

BALANCE, RESET

ANNUAL PARAMETERS

Tno

CALCULATE ANY SNOW,
ADD TO SNOWPACK, OR
CALCULATE SNOWMELT

CALCULATE SOIL WATER
MOVEMENT, ET, SEEPAGE

CALCULATE RAINFALL
EXCESS, PEAK FLOW,
El ASSIGN EVENT ID

RESET DAILY LEAF
AREA INDEX VALUES

Figure 1 1 -5 . — Generalized flow chart for HYDTWO (hydrology option 2)

172

Table 1 1-5. — Hydrology model parameters

Model
option

Reference/definition

Source of estimate

Quality

DACRE Both

RC Both

Field area in acres.

Saturated hydraulic
conductivity, in/hr
(Ks in equation 1-9).

FUL

--Both

Portion of plant-avail-
able water storage
filled at field capa-
city.

BST

—Both

Portion of plant-avail-
able water storage
filled when simulation
begins.

CONA

—Both

Soil evaporation param-
eter, as (eq. 1-43).

POROS—

--Both

t, soil porosity (eq.
1-8).

BR15

—Both

Immobile soil water con-
tent.

TEMPO--

—Both

Average monthly temper-
ature (read values) °F.

RADIO-

—Both

Average monthly net radia-
tion (read 12 values)
langleys/day.

—Both

Winter cover factor (1 for
crops, 0.5 for grass.)

X(I)— -

—Both

Leaf area index, day I
[must specify X(l) and
X(366)].

Initial abstraction co-

efficient CN method

(eq. 1-2).

Channel slope (CS in eq.

1-7).

SCS curve number for AMC

condition II.

Watershed length/width

ratio.

Plant-available water

storage in 7 soil lay-

ers, in inches.

.—2

Depth of surface soil
layer.

..._2

Depth of root soil zone

GA--

.—2

G in equation 1-16.
Effective capillary
tension.

RMN

— _2

Manning roughness for
field surface (Cc in
eq. 1-30).

SLOPE

..._2

Average field slope (Sc
in eq. 1-36).

XLP

.—2

Length of flow plane

(L in eq. 1-36). j

Measurable

Good.

Estimate from SCS
soil class; or
measure, infil-
trometer or in
lab.

Poor to good;
sensitive.

Estimate or from
reference.

Well defined
quantity.

Field measure or
estimate.

Not sensitive.

Estimate from hand-
book.

Fair.

Estimate or measure.

Not sensitive.

Estimate or measure,
or from (1).

Not sensitive.

Climatological data
Climatological data

Good, but only

average.
Good, but only

average.

Crop information

Rough.

Crop information
handbook (table
II-8).

Good.

Use 0.2s in absence
of measured value
(5).

Fair.

Field measurement

Good.

Handbook; soils data.

Fair.

Watershed map

Good.

Difference between
total soil porosi-
ty and 15 bar water
content.

Fair to good.

User estimate

Knowledge of soil ;

rooting depth.
Soil data; infil-

trometer tests.

Varies; sub-
jective.
Fair.

Fair to good.

Handbooks; field
observation.

Good; sub-
jective.

Maps; field survey.

Good.

Maps; field survey.

Good.

173

Table 1 1-6. — Hydrology parameters input file

Both Options

Card 1-3. TITLE

TITLE Three lines of 80 characters each for alphanumeric
information to be printed at the beginning of the out-
put, format (20A4)

Card 4. BDATE, FLGOUT, FLGPAS, FLGOPT, ELGPRE

BDATE The beginning date for simulation. It must be less
than the first storm date (Julian date) , e.g. 73138

FLGOUT 0 for annual summary output

1 for storm by storm and annual summary output

FLGPAS 0 if no hydrology file is to be created

1 if the program should create a hydrology file for use
by the Erosion Program

FLGOPT' 1 for the daily rainfall model (option 1)

2 for the breakpoint or hourly rainfall model (option
2)

FLGPRE 0 for breakpoint precipitation data
1 for hourly precipitation data
(only used for hydrology model option 2)

Card 5. EACRE, RC, FUL, BST, CONA, POROS, BR15

DACRE Field area (acres), e.g. 3.2

RC Effective saturated conductivity of the soil (in/hr),
e.g. 0.19

FUL Fraction of pore space filled at field capacity, e.g.
0.75

BST Eraction of plant-available water storage filled when
simulation begins, e.g. 0.50

CONA Soil evaporation parameter, e.g. 3.75

POROS Soil porosity (cc/cc) , e.g. 0.41

BR15 Immobile soil water content at 15 bars tension (in/in),
e.g. 0.17

174

Table 1 1-6. — Hydrology parameters input file— Continued

For Daily Rainfall Model (option 1)

Card 6. SIA, CN2, ChS, VvDa, RD

SIA Initial abstraction coefficient for SCS Curve Number

method, e.g. 0.2

CN2 Two condition SCS Curve Number, e.g. 80.0

CHS Channel slope, e.g. 0.022

VvLW Watershed length/width ratio, e.g. 2.1

RD Maximum rooting depth (in) , e.g. 24.0

Card 7. UL(l-7)

UL() Plant-available soil water storage for each of 7 soil

storages (in), e.g. 0.16

(Top storage depth=l/36, 2nd storage depth=5/36, other
storage depths=l/6 of rooting depth (RD, card 6))

por Breakpoint Rainfall Model (option 2)

Card 6. DS, DP, GA, RMN, SLOPE, XLP

DS Depth of surface soil layer (in), e.g. 2.0

DP Depth of maximum root growth layer (in), e.g. 22.0

GA Effective capillary tension of soil (in), e.g. 13.0

RMN Manning's n for overland flow, e.g. 0.03

SLOPE Effective hydrologic slope (ft/ft), e.g. 0.015

XLP Effective hydrologic slope length (ft), e.g. 350.0

Both Options Continue

Card 0,9. TEMP(1-12)

TEMPO Average monthly temperatures (degrees f .) , e.g. 45.0

Card 10,11. RADI(1-12)

RADIO Average monthly solar radiation values ( lang leys/day) ,
e.g. 218.0

175

Table 1 1-6 — Hydrology parameters input file—Continued

Card 12. GR

GR Winter cover factor

1.0 for crops
0.5 for grass

Card 13. LDATE,AREA

LEATE Date (Julian day) , e.g. 001

AREA Leaf area index for the crop grown the first year of

simulation, e.g. 0.0

A card 13 is repeated as many times as is necessary to define the
LAI curve. The first card 13 should always have the date 001.
The last should always have the date 366.

Temperatures, solar radiation values, and leaf area index parameters can
be updated at the end of each year. If they're to be updated, they will be
read in the same sequence and format as the initial inputs. The winter cover
factor (GR, card 12) will be read if the leaf area index is updated.

Card 14. NEWT, NEWR, NEWL

NEWT 0 use the temperatures from last year

1 read a new set of temperatures
[-] stop program execution

NEWR 0 use the solar radiation values from last year

1 read new solar radiation values

NEWL 0 use the leaf area index from last year

1 read a new set of leaf area index values and process
them

A card 14 is read after each year of simulation, lb stop execution of
the program a negative value is read in NEWT. If any of the "NEW" parameters
call for further input the appropriate data must be inserted after that card
14. That is cards 8 and 9 for NEWT, 10 and 11 for NEWR, and card 12 and a set
of 13's for NEWL.

176

Table 1 1-6 . — Hydrology parameters input file--Conti nued

The following sample is a complete data set, good for a three year run
using the daily rainfall option.

CARD
NO HYDROLOGY PARAMETER DATA

1 DAILY HYDROLOGY PARAMETERS - GEORGIA PIEDMONT

2 MANAGEMENT PRACTICE ONE

3 CONTINOUS CORN - CONUENTIONAL TILLAGE
4
5
G
7
8 45.0 47.0 52.0 G1.0 70.0 77.0 79.0 78.0 73.0 G3.0

9
10
11
12
13
13
13
13
13
13
13
13
13
13
13
14
14
14

73138

3.200

0.190

0.750

0.500

3.750

0.410

0.170

0.200

80.000

0.022

2.100

24.000

O.1G0

0.820

0.720

0.520

0.G10

0.700

0.GG0

45.0

47.0

52.0

G1.0

70.0

77.0

79.0

51.0

44.0

218.0

290.0

380.0

488.0

533.0

5G2.0

532.0

268. 0

211.0

1.000

0.000

122

0.000

152

0.200

1GG

0.200

183

1.000

192

2.500

197

2. GOO

202

2-700

228

2.200

255

0.000

3GG

0.000

-1

508.0 41G.0 344.0

contained within the same columns, and the decimal point must be entered in the
number. If the example has a decimal in it, the parameter is real; otherwise,
it is an integer. The alphanumeric input is read with A4 formats. Specific
instructions are given whenever alphanumeric input is required. Sample deck
representations are shown schematically in figures 1 1 -6 and 1 1 -7 for hydrology
model options 1 and 2, respectively.

Climatic Data

Monthly mean air temperature and mean daily solar radiation data are re-
quired inputs used to calculate daily evapotranspiration. Daily values of
temperature and radiation are calculated from the mean monthly values fitted to
an annual curve by Fourier analysis ( 2J . The user can use long-term averages
or actual monthly data for the specific period of simulation. Temperature data
are published regularly by the NWS ( 7) . Current solar radiation data are not
readily available. Daily and monthly data were published for selected loca-
tions from 1954 through 1973 (8). Publication was suspended in 1974, and only
selected stations were published for the entire United States after that time.

177

The user can obtain monthly average daily radiation data from the Climatic
Atlas of the United States (6). A summary of monthly radiation data is shown
in table 1 1-7 for the user's convenience.

LDATEN AREAN h

LBATE2 AREA2

L'Ll

RADI2 RADI3 RADI4 RADI5 RADI6 RADI7 RAD 1 8 RADI9 RAD 1 10 I

TEMPI 2 L

'2 TEMP3 TEMP4 TEMP5 TEf;P6 TEti?7 TENP8 TEMP9 TENPlp'l

SI A CM2 WLW

DACRE RC FUL EST CONA PDRDS
BDATE FLGDUT FLGPAS FLGDPT FLGPRE

THE FIRST THREE CARDS ARE USED FDR IDENTIFYING
INFORMATION DN THE RUN (E.G. LOCATION, PRACTICES) ETC.) FDRMAK20A4)
(HYDROLOGY OPTION ONE PARAMETERS FILE)

TITLE(I,J) 1 = 1 TO 3. J=l TO 2C

Figure 1 1-6. — Sample deck arrangement of input parameters for hydrology model

option 1 .

LDATEN AREAN

LDATE2 AREA2 |

LDATE1 AREA1 ' 1

GR I

RADI11 RADI12

RADII RAD 12 RAD 1 3 RAD 1 4 RAD 1 5 RAD 1 6 RAD 1 7 RAD 1 8 RAD 1 9 RADII 0 f
TEMP 11 TEMP 12 \

RMN SLOPE

TENP9 TEMPI 0 |

BDATE FLGOUT FLGPAS FLGOPT FLGPRE

THE FIRST THREE CARDS ARE USED FOR IDENTIFYING
INFORMATION DN THE RUN (E.G. LOCATION, PRACTICES, ETC.) FORMAT(20A4)
(HYDROLOGY OPTION TWO PARAMETERS FILE)

1=1 TO 3, J=l TO 20

Figure 1 1-7. — Sample deck arrangement of input parameters for hydrology model

option 2.

Parameter Estimation

Daily Rainfall Model (Option 1)

Most parameters are easy to evaluate from existing data using the proce-
dures outlined in table 1 1-5. A slightly expanded explanation follows some

178

procedures for estimation. The beginning soil water storage, BST, is generally
unknown because most simulations begin several years in advance, and few loca-
tions measure soil water. Since yearly variations in soil water are usually
small on January 1, BST can be estimated adequately for most areas. Since this
estimate only affects model results before the first filling of soil storages,
it is not very important except for short simulations (1 or 2 yr) or low-rain-
fall areas.

The winter cover factor, GR, reduces soil evaporation as a result of such
ground cover as dormant pastures or heavy crop residue. The value of GR varies
from 0.5 for excellent cover to 1.0 for bare soil.

The CN initial abstraction parameter, SIA, normally equals 0.2. The user
may assign SIA any value grester than zero and less than one, however, for un-
usual applications. The 0.2 value generally is recommended unless justifica-
tion is strong for another value.

Water storage, UL, is calculated for seven soil layers. Thickness of the
layers normally is selected so that their sum equals maximum root depth.
Thickness of the top layer is 1/36 of the maximum depth, the second layer is
5/36, and the remaining five layers are 1/6 each.

Leaf area index, X(I) is used in both model options. Table 1 1 -8 gives
some typical leaf area index distributions for normalized times through a grow-
ing season for several crops. These values must be apportioned between actual
local planting and harvesting dates. The distribution is specified as shown in
figure 1 1 -8. Points for day = 1 and day = 366 are necessary.

The watershed length-width ratio is used in predicting peak runoff rates.
Length is determined by measuring the distance from the watershed outlet along
the main channel to the most distant point on the watershed boundary. The
length-width ratio is computed by squaring length and dividing by the watershed
area.

Breakpoint Infiltration Model (Option 2)

The breakpoint infiltration model (option 2) uses two parameters and a
variable, as does the CN method, but can incorporate any additional data di-
rectly into its parameters. Any inf iltrometer test can be used, for example,
directly to yield values for parameters GA and RC.

Variable D is a straightforward estimate of porosity available in the soil
surface at the beginning of the storm. Using Sw for relative saturation, Sw =
0 for air dry, and Sw = 1 for totally wet conditions, and i for porosity,

D = i (1-SW) . CII-1]

An AMC-III condition is analagous to a large value of D (D=*0. 3-0.4), AMC-I con-
dition is similar to a small value of D (D=0. 05-0.1) . Water contents are as-
sumed not to dry below the BR15 value. Water contents are assumed not to dry
below the BR15 value. Values of this parameter may be estimated from informa-
tion such as Holtan and others (1).

179

ooccocor-icrv r^ lo i

irnc CO ID « CO IT. LC LD

«a- ld co «a- «=j- ^i- «* co «a- ■

CT> CM CM C O CT : i-H I

OhhCWW
OC\*MnCO
CO CV CM CM CV i— I

O O O O CT.

LD LO HOI

O ID i-H

>-i c o en .— n-

HCOOl t-Hi-H LD LT i— < O

•-h O O l
HW CM

CM CM CO CM CM .

cd cm

r- o

LO CM

OHCCOOl i— < i— * LT ID rH O

8^ s

r^ id co

or-- cd

CM CM CM

SRI

CO CM CD IC^O

IT) O CC .-i LO r-

CM CM .— I ID LO LT)

CM ^J- LO '

«d- in «s- I

CTi O CJl OCfi o i— i

IOOCOOM ID IT)

OCl CT) ID
O CM CO 00

O CM O CD

CD O CO CD CT) O .—i

ID CO CO
CT) .-i CO
ID ID ID

cm ■— i CD «d- ■— icocOf-i

.— < CM CM LO CO .— It- ICO
^DlDr^<tin(0^lO

CT) ID LD O
CO CD O «3"

i— i t— ' CO ID *3-

rHlOCOOO)

ID ID ^ ID «3"

«H»

cowiHCvjoiai'-ifo

t-H CO i-H CM i-l

ID CO CO CT)

CT) i— i O CD CT)

CD 00 LO
CO ID CM
ID ID ID

CMCMCMr-t-lCOOO

cocor^cD«d-c-coLo

IDlONtlOlOlON

O ID CD ID
CM r^ O CT)

inio io>a-

CT) CO CO CM "3-
CM 00 CO CO CO

co o •-*

CD >* CM CT) I

ID O >* H
CT) CO CT) CD
ID ID LO ID

ID CO CO CT)

ID CO CM O
CM CD CO -—I
LD r-- LO ID

CD o CD O CD

r~. lo cm co i

■— ( ID ID

IT)ID ID CM 00 CO

rHIO CO CD ID ID

ID ID ID ID ^t" ID

CM ID CM CO LD ID

N LOLOtOlO^O

CM p- CO CO CD CO ID

id id ocorvino)

ID ^T ID «* ID ID LD

2S 3 §

CM .-H

.-1 r-i O O CD O '

LO *J" LD LD LO LO

CD O CM CD CO
ID CO r^ O CD

CD CT) LO CM
LD CM CO LO
LO ID «^ «*■

O CD CM CT) CO

LO CM

ro 00

CM CM

LO LO CD
CM CM ^3"
LO LO LO

OtvOlrvOCONt

CT)CMLOCT)p-~r~~cor-.
co«d-LOco*J-«3-"=rro

■-I "3- CO CM
O CO CM CM

LD CD

CO ■— 1 LO

r-.-i.-iooooO'-'co

«3" CTl

r-~ CD

LT> CO
I-H O

CM CD rH
CO O CT)
CO «3- CO

LO
CM

NO)CMC\Jr- ir-LDLO
IflCOHOfOlOiJH
CMCM«a-COCOCOCOCO

00 «=*-

LD CM

cm ro

M CM

LO CT)

CM .-I ID

COrtHCflOCOrH*

ID CD

LO CT!

co co

LO CO

CD •— I LO
O O .-1

«3->*lD«3-00LOfOrO
sooo<t<f Mccg

r-Hf-lCOCMCMCMCMCM

O CM
CM CM

CM CO

.—( LO
CM .-1

CM hni

O ■* *3-

o lo Lo r-

00 CO — i CO
CO CO LO «d"

O O O CD CO

r-< CD

>d- «d-

en o

CT) LO

r-~ o co lo .—i
co co co *3- co

O LD
CD CD
CM CM

CM O
CM O

*3- «3-

LO O

co «j-

CM CM

r~- cm

O O .-1 O CD

i-H CD

"=3" LO

o o

.-H CM

CT) LO

co r» r- CD r~-

00^-

CO CTl

CO CO

LD O

i-H CM i-H

«* r-H If)

rt CM CM CM

CM CM CM CO CO

O < i— cc

</) OO 4-> O 1-

U Z ■!-> U- >

^ <: •■- i— i ro

h- i^ _1 _l D

t. c ro o -i-

T3 TD '

: •— <=
: 3 ro
i o t-
I CO CD I

i.— r- cEa..— ro4__oj-
. ro ■— -i- ro Eoi— •!-•—• c roo
: o_ cu ro •— <ocj:+jl<oq)i
i<tcooz:i— 0<03ia<

l UJ «£ O

c «t ro

co u o z •
•J == •= I

c< En o
<. z o

180

CM 00 CM CM LD

O Ol LD oc id

CO CM CO CM CO

00 ID
I CM .-<,-< .-H .-< ^1 •-<

«3- 00 LO CM' LO CTl C
CM IT) .— I ■— ' r-» ID T-~
■-1 ,-1 .-i .-I ri ,-H CMCM

8* 2^

en c oo oo c ic
its Hin «a- r~-

.-i CM CO

LO LO '

O 00

I CM CM r-l ■-! r^t ,-. ■

Hn«t <— i d cc lo «a- i-icti

Mrv<J CM0O ID ID CO LO LO

O LO LO HN LO CO LD LD »-H

** CO CO CMCM CM CM CM CMCM

lo lo o

cn «a-
cr> o

■-( CM

Oi 00
CO •-<
CM CO

ONtCOOM
CM ^t- 00 CM 00

CO CM
«3- CO
CM CM

00 O CM

o <-<

CM ID C 00 LO

o <-i

r- «a- id

CM CM CM

LO 00
CTl CTi
CM CM

CT1 CM

i-< CM CTl 00 CO

co >* oo <— i cti

CM CM CM CM CM

CM 00

cm lo

CO CO

CM
CM

2~ 8^ SS

rtWC
CO CO .-H

.-H ,-1 CM

O CO
LO LT)

«3" ID CO CO

r- OCOCTi
ld io irrn-

.-I CM CO

ON.lOtN

r-H ID CTl CO

.— t CM CO -— 1

O 00 IDN^f •-< O

CTl O ID
00 LD 00
MCOCO

ID >* '

§5:

CO LD O CTl CO

ai ld oo co

ID .—I CO CM

co «d- co «=r

ID T— CM .-no

«tf- CM
LD LD
CO CO

CTl O LD

.-i ai o

CO CM CO

CTl LO
O LD
CO CO

LD LD CO
00 LO LD
CO CO CM

O LD
LD LO
CO CO

<Tl 00

r- r-~ co

.-H CTl

00 CTl CM

i— I ID

C0«t oo

O ID

.-1

00 CM CM

ai en

LD CTl CM
O LD CTl
CO CM CM

.— 1 ID

co co

CM CM

CM CTl .—i
CM .—I CM

O LD
■— I CM
CM CM

.-H «3- CO

O0 !-~

r-- LO CO

CM r-H .-1

o o

LD U- CM

«3" i-i CO

CO CM
CO LD

co en o

ID CM ^1-

•-< o

CM CO

CTl «3"
CO 00
CO CO

CM CM —l

.— < CM CO >— I

.—I CM CD LD CO Ol

rt<J-rtOlCO

CM CO .-h

cn en o cm

2-i

z s: z .

4-> <c

OIZ

S- 1- <

it OlO
OQ.W

I . — • t-

i -t- <o

'DC C 3 Z= .

O ct

(1J +J 4-> O
3 l/> V) >— i

•— o <o 3:

maiLuu

c o

cn i

:o s.

a; +j

•»-> o ■— cc

-l-> .— OO <_> 3

1-4 _ Q

LDoOOi— iID^fOTJi — CJ

:cn+j.1-i1iOj= i>ce!-xo-ct:<j!--^cc

: W <t> g 00 CJ +-> "=C UJiJllJ30''-P><llO

:««l<CI.D>1Wr)llJ2Il>.£C >,JZ ■

. i— s- 3tt:-r- o «t

: C3 CO OO 0Q

CD i— +J 0J <0 (_> O.

>Ld_J3(/13<2'-|UV)W

-c aj +-> 3:
P s. m|_.

3 aoi aca

181

c. cm to m

CM in *t CM IC LO lO CM to "3" O C Cn CM CM C « co cc
Cn LO to «3- CO *r to «* CM Cn in CM cn r~- r^ o cm ^1-0
co r,m «j- if) «3 q- i1 «a- co co co n w con co ^-^a-

Cn i-h C ^OiCCI

in 01 a ocr. cn o co r-.cc cn in
hmo «3- c r-. oo cn r~. i-h co cm

CM CO CM CM CM

IT IC OO in O Cn IC

r-l^lDN O ■-
CM rH I-H l-H

Ol IT, O IQ Cn «3 lO
in IOIC i— I

en 10 i-i

On cn oc «- .-h lO o

CM i-H l-H ^Hr-H -H ,-H i-H ,-H

cn to to «*• oo n

00 O C i-i

CM CM CM ~

c ^h^h,-hc»o co cc cm o «* cn

cm cm co co co i

en co vo

lO CM lO LO in CM ^J- CM O0 lO CO lO ^f LT) Cn Cn i-H

• ONOl to o o o n ^ <to «d- COCM ,-H

CM CM CM CM i— i i-H i-H,— i ,-H CMCM *3"

CVi l-H I-I

Oin r-~

O CTi ^H ,-H ,

,-H O ^H

COCOCMO «3" cn ino
CM lO --h «» «3; C m 00

■-H «g- «a-

,-H cn ,-h

*3- .-h ioo

CM ,-H .— I l-H ,-H

l-H l-H ,-H l-H Cn O

£g

«J- CO CO

in cr> ■—< cm *t

lOO "3" •-< rH

CM C\I r-4 t-l

cn<t in ,-h ,-h

LO lO in in in

CM CM ,-H O ID

to «* to ,-h cn

in in in to in

O ID
CO CM

imo ^ ^

to cn to ,-h
cn cn to co
«■ to in in

cn ,-h ,-h co cn

in to «3- in

in in HMitirtH
in in to r- to to to

to to in in

10 <j win

in r^ in in in

«3- 10 in

in in in in

10 t-» .3- cn
10 oo in ,-h

CO ,-H ,-H Cnrt ,-H ,-H ,-H I

in cn cm cn r~- ■— i (

CM Cn T CC i-h«3-CMCO «3-

CM r-. ,-h .— < oo to to o cn
in«-«3-«a- moo«a-*3- co

>-H ^H Cn l-H

o o i-H en en

co r- cm
o en o

CO CM CO

cn o
o cn
M- CO

to in

rH cn
OvJ oo

ior~.cn

O 00

£j

in cn

co o *r
00 o o

-h CM CM

cn cn

CO CM

cn ^h
to O
,-H oo

to r-~ o

o oo

<*•

to en

in oo io

CM CM CM

i-i in

in o

CM CM

o cn

«3" CO

on co

CM rH) r-H

,-H CM ,-H ,-H

oo cn

CM CO

oo oo

l-H ,-H ,-H CM ,-H

r-H O O

z: i i ii qj i i

_j o i— j-> cua) .— ij=

i— _o c «a; o

cn e •- s: jc

O <a

o i_> ,— o -o oo

Q. i_ •■- 00 .

fo pc O-uj

CDi— +-> 'SI f— ■■- O 4-> I- TJW 4J TOLlJ J J

«uoo iooouj<cos:zo.iooz=3<->r5Q;z

r— CTl -r- +J

t i— -o > o i- ■

I -r- .r- l/l in O

I > CC <= <D 3

«3- co «3- i-i in

CO CO O0 CO CM

<U +j I (_
cn-i- O

LO 1-J

cm to in cn

CM CM O CO

CM i-H CM i-H

1 Ol tCl

i-H CM CM

0) i. jz ro i— >— i o

zoxtoujii.i:w<

, ,_ . .,_ _ o

■HJ+Ji-lL

. to z: u_ oo o_

i- -r-

ai E

00 r— h-

182

Table 1 1-8.— Typical leaf area index distributions for crops

Portion of

Leaf area index-

growi ng
season

Corn

Cotton

Sorghum

Oats

Wheat

2/
Pasture^7

Barley

Soybeans

0.0

0.00

.09

.13

.09

.42

.47

1.84

.44

.15

.19

.28

.19

.84

.90

3.00

.88

.40

.23

.50

.23

.90

3.00

.90

2.18

.49

2.14

.54

.90

3.00

.90

2.97

1.16

2.96

1.35

.98

.90

3.00

1.58

3.00

2.97

3.00

2.98

2.62

1.62

3.00

2.96

3.00

2.96

3.00

2.70

3.00

2.92

2.72

2.92

2.72

3.00

1.96

3.00

2.30

1.83

1.78

1.84

3.00

.96

2.14

1.15

2.00

1.0

.00

1.00

.00

.50

.00

.50

1/ Good production assumed for all crops. LAI should be lowered for poor
production.

2/ No grazing assumed. LAI must be lowered if grazed or mowed according
to height of plants.

-2.0

/ \ CORN, OHIO

IxJ

MEADOW GRASS, /

/ v \^

Li.

< 1.0

TEXAS /

7 ^ \

^•^^ /

\^^ /

/ \ \

/ H

/ ^--^*"

— K

1 4-1. *-Cl

i 4 \ • '

IOO

150 200

DAY OF YEAR

250 300

350

Figure 1 1-8. — Leaf area index for meadow grass at a location in Texas and
for corn at a location in Ohio.

GA is a parameter characteristic of the soil type and composition. Table
1 1-9 shows how GA may be estimated in relation to the SCS hydrologic soil
groups.

RC represents effective saturated hydraulic conductivity. Experiments and

183

Table II-9. — Parameter estimation of the infiltration model —

G (in)

Hydrologic soil group
A B C D

RC (in/hr;

Hydrologic soil group
A B CD

Expected range
of values

3-6 7-11 12-17 18-22 0.3-3.0 0.1-1.0 0.05-0.5 0.01-0.2

Mean for
estimation

Land use^ W™]0*]*
condition

Row crops:
Straight

Poor
Good

0.4
.45

,18
.21

0.06
.09

0.03
.04

Contoured

Fallow

Poor
Good

.25

,12

.1
.12

.04

.05

.07

.02

Small grains
and meadow:

Straight

Contoured

Poor
Good

.05

.22

.09

.04

.43

.12

.07

.05

.25

.07

,65

.14

.09

.08

.04

.42

.12

Range/pas-
ture:

Straight

Poor
Good

Contoured

Poor
Good

.75
2.0

.35
.7

.13
.24

.04
.12

Meadow

Good

1.2

.46

.23

.13

Woods

Poor
Good

.75

1.5

.36

.50

.17
.24

.09
.14

1/ Tentative.

2/ SCS presentation of CN for terraced systems is omitted since this af-
fects routing, not condition of soil.

3/ Values given will reproduce CN estimates for a storm of 4.0 in covering
4-hr duration, with mean value of G.

theory suggest the approximation that variations of this parameter can relate
changes to soil condition. Thus, table 1 1-9 shows a proposed guide to use in-
stead of table 9.1 of NEH4 ( 5J . Although the variations shown for RC as a
function of soil cover complex are tentative, they are useful in a relative
sense. A better version of this table requires use of the data from which the
SCS table was developed. Actual value for RC to reproduce the runoff predicted
by the CN method depends on the depth and duration of the storm from which the
tabulated value of CN was obtained. Should such data be available, as are much
SEA-AR data for any model application, they can be used to improve parameter

184

estimates. Notable exceptions to these RC values include a cracked clay, soil
class D, which will exhibit apparent large RC early in a storm, confounding
predictive accuracy.

The parameters RMN, SLOPE, and XLP are used to estimate peak rates of run-
off. Effective length of flow in the field is determined by estimating the
median flow path, including a fraction of concentrated flow path length. The
fraction of concentrated flow path length to include will be large if the flow
is through a rough or mildly sloping channel.

If the length of the watershed flow path is broken into N regions of dif-
ferent slope and roughness, equivalent single-plane values of RMN and SLOPE can
be determined. For each segment or subplane, j, j = 1, N
equation [1-25], volume I, chapter 2, defined as

there is an aj as in

= Ci /Si

where C and S are roughness and slope, respectively,
equation [1-34] is

[II-2]
From volume I, chapter 2,

Cc ^c " ac

2"" (aj)1/m
-j=l

CII-3]

for x0 = 0 to X|\| = L where Cc is composite roughness, Sc is composite slope, L
is horizontal length of slope, and b = (m + l)/m. If Sc is taken as the equiva-
lent single slope (SLOPE),

- - ^c

c = Vs"c

[II-4]

and the equivalent roughness, RMN, is 1 .49/Cc for the Manning flow equation.

OUTPUT

Hydrology output is composed of input information and calculated values.
Sample input information included in the output is shown in figures 1 1 -9 and
11-10 for options 1 and 2, respectively. Daily and annual simulated output
data are the same for both options. Sample output data are shown in figure II-
11. These data are transmitted to the erosion model in the hydrology pass
file. Figure 11-12 is a sample of averages and statistics calculated for the
period of simulation. Output for the simulation period also includes monthly
totals and means of rainfall, runoff, ET, percolation, and average soil water
(fig. 11-13). Data include annual totals of each component.

185

HYDROLOGY OPTION ONE

(DAILY PRECIPITATION UALUES)

DAILY HYDROLOGY PARAMETERS - GEORGIA PIEDMONT

MANAGEMENT PRACTICE ONE

CONTINOUS CORN - CONVENTIONAL TILLAGE

43.43

79.90

236.29

555 . 38

0.1B0

0.080

MONTHLY MEAN TEMPERATURES,
45. B4 52.15
77. S3 71.18

MONTHLY MEAN RADIATION,
291.73 375.07
439.93 41G.B0

DEGREES FAHRENHEIT
61.21 70.39
62.12 52.94

LANGLEYS PER DAY
463.97 534.61
327.70 257.05

77.23

46.10

568.07
223.59

LEAF AREA INDEX TA3LE
DATE LAI

0.00

122

0.00

152

0.20

166

0.20

183

1.00

192

2.50

197

2.60

202

2.70

228

2.20

255

0.00

3EB

0.00

WINTER C FACTOR = 1.00
LAI-DAYS = 151.15

FIELD AREA = 3.200 ACRES

ROOTING DEPTH = 24.000 IN

RETENTION RATE = 0.190 IN/HR

FIELD CAPACITY = 0.750

INITIAL STORAGE FRACTION = 0.500

INITIAL ABSTRACTION = 0.200

EUAPORATION COEFFICIENT = 3.750

SCS CURUE NUMBER = 80.000

CHANNEL SLOPE = 0.022

WATERSHED LEN/WIDTH RATIO = 2.100

PEAK FLOW RATE COEFFICIENT = 9.087

PEAK FLOW RATE EXPONENT = 0.840

UPPER LIMIT OF STORAGE = 4.190 IN

IMMOBILE SOIL WATER CONTENT = 0.235 IN/IN

INITIAL SOIL WATER STORAGE = 2.095 IN

0.820

UPPER LIMIT OF STORAGES

0.720 0.520 0.610 0.700

INITIAL STORAGE
0.410 0.360 0.260 0.305 0.350

0.660
0.330

Figure 1 1-9. — Sample output of input values for hydrology option 1

i86

HYDROLOGY OPTION TWO

(BREAKPOINT OR HOURLY PRECIPITATION VALUES)

BREAKPOINT HYDROLOGY PARAMETERS - GEORGIA PIEDMONT

MANAGEMENT PRACTICE ONE

CONTINOUS CORN - CONVENTIONAL TILLAGE

MONTHLY MEAN TEMPERATURES»
43.43 45. G4 52.15
79.90 77. G9 71.18

MONTHLY MEAN RADIATION,
236.29 291.73 375.07
555.38 439.93 41G.G0

DEGREES FAHRENHEIT
61,21 70.39
62.12 52.34

LANGLEYS PER DAY
4G3.97 534. Gl
327.70 257.05

LEAF AREA INDEX TABLE
DATE LAI

0.00

122

0.00

152

0.20

166

0.20

183

1.00

192

2.50

197

2.60

202

2.70

228

2.20

255

0.00

366

0.00

77.23
46.10

568.07
223.59

WINTER C FACTOR = 1.00
LAI-DAYS = 151.15

EFFECTIVE HYDROLOGIC LENGTH
EFFECTIUE HYDROLOGIC SLOPE
EFFECTIUE MANNINGS N
DEPTH OF SURFACE LAYER
DEPTH OF REMAINING ROOT ZONE
EFFECTIUE CAPILLARY TENSION
EVAPORATION COEFFICIENT
SAT. CONDUCTIUITY CULTIVATED
SAT. CONDUCTIUITY FALLOW
SOIL POROSITY

IMMOBILE SOIL WATER CONTENT
UPPER LIMIT OF STORAGE
INITIAL SURFACE STORAGE
INITIAL REMAINING STORAGE
TOTAL INITIAL STORAGE

150.000

0.015

0.030

2.000

22.000

13.000

3.750

0.190

IN/HR

0.152

IN/HR

0.410

0.170

IN/IN

4.440

0.240

1.980

2.220

Figure 11-10. — Sample output of input values for hydrology option 2

187

DATE

RAINFALL

RUNOFF

PERCOL.

AUERAGE

ACTUAL

POTENT.

TEMP.

SOIL W.

JULIAN

INCHES

DIG. F.

IN. /IN.

INCHES

74001

0.1100

0.0000

44.0823

0.3242

0.0000

74002

0.1G00

0.0000

44.0015

0.3278

O.OGOO

0.0000

74003

0.3700

0.0000

43.9259

0.3401

0.0000

74004

C.1700

0.0000

43.8555

0.3441

0.0000

74007

0.3400

0.0000

43.7323

0.3426

0.0000

74011

0.0800

0.0000

43.5660

0.3419

0.0000

74020

0.8700

0.0000

43.4426

0.3320

0.0000

74024

0.2400

0.0000

43.5163

0.3500

0.0000

74023

0.1000

0.0000

43.6840

0.3453

0.0000

74029

0.2G00

0.0000

43.8253

C.3515

0.0000

74037

1.7000

0.2584

0.5470

44.1300

0.3428

0.0000

74038

0.2000

0.0000

0.1085

44. 629 0

0.3664

0.0000

74045

0.G500

0.0000

0.0823

45.1264

0.3558

0.0000

74046

0.9000

0.1765

0.6243

^5.6823

0.3664

0.0000

74050

0.1G00

0.0000

46.0771

0.3532

0.0000

74053

0.5000

0.0000

0.0132

46.6726

0.3573

0.0000

74078

0.1500

0.0000

43.6677

0.3397

0.0000

74080

0.6500

0.0000

53.0355

0.3402

0.0000

74084

0.3500

0.0000

53.8765

0.3487

0.0000

7408S

G.0800

0.0000

54.7343

0.3513

0.0000

74088

0.6900

0.0000

0.1064

55.3181

0.3577

0.0000

74094

1.3000

0.1433

0.53S5

56.50SG

0.3547

0.0000

74102

0.0500

0.0000

58.6436

0.3501

0.0000

74103

0.9500

0.0357

0.2315

60.0387

0.3664

0.0000

74112

0.3000

0.0000

61.6025

0.3503

0.0000

74122

0.0900

0.0000

64.561'j

0 . 3425

0.0005

74124

0.3500

0.0000

66.4012

0.3413

0.0030

74125

0.7400

0.0000

66.8540

0.3650

C.0050

0.0050

74131

0. 1000

0.0000

67.8323

0.3515

0.0285

74132

0.5000

0.0000

63.9134

0.3536

0.0343

74135

0.1000

0.0000

63.4830

0.3519

0.0553

74143

0.2600

0.0000

70.3953

0.3428

0.1378

74144

2.500C

0.S378

0.9712

72.1848

0.3664

0.1509

7414G

0.2800

0.0000

72.5647

0.3595

0.1791

74151

0.5000

0.0000

73.4205

0.3523

0.2608

74159

0.3000

0.0000

74.8939

0.3433

0.4044

741G1

0.2500

0.0000

75.5311

0.3376

0.4403

74171

0.4800

0.0000

77.0056

0,3279

0.7179

0.7173

74178

4.2600

1.2540

1.2569

78.2533

0.3221

1.1551

1.1581

74198

0.1100

0.0000

79.4839

0.2880

3.5846

4.7379

74204

0.1100

0.0000

79.8335

0.2358

3.6409

6.2023

74205

0.5800

0.0000

79.8347

0.2520

3.8284

6.4397

74207

0.5100

0.0000

79.735-4

0.2506

4.1690

6.9052

74208

2.8400

0.5856

0.0000

79.7254

0.3385

4.3985

7.1353

ANNUAL TOTALS FOR 1974

PRECIPITATION
PREDICTED RUNOFF
DEEP PERCOLATION
TOTAL ET

BEGIN SOIL WATER
FINAL SOIL MATER
WATER BUDGET BAL.

40.260
3.516
4.316

30.929
2.095
2. 994
0.000

Figure 11-11. — Sample output of daily data and annual summary
from the hydrology model. Full year of daily values not
shown.

188

AUERAGE ANNUAL UALUES

PRECIPITATION = 44.255

PREDICTED RUNOFP = 5.513

DEEP PERCOLATION = 5.372

TOTAL ET = E3.7S5

AUG. AUAL. STORAGE = 2.270 IN

FINAL AUAL. STORAGE = 2.881 IN

FINAL STORAGE FOR EACH FRACTION
0.000 0.000 0.000 0.113 0.458 0.525 0.213

MINIMUM TOTAL STORAGE WAS 0.000 ON 74200
MAXIMUM TOTAL STORAGE WAS 3.143 ON 74144

Figure 1 1-12.— Sample output of averages and statistics
calculated for the period of simulation.

HYDROLOGY SUMMARY

DAILY HYDROLOGY PARAMETERS - GEORGIA PIEDMONT

MANAGEMENT PRACTICE ONE

CONTINOUS CORN - CONVENTIONAL TILLAGE

1374
MONTH RAIN RUNOFF ET PERC AUG SW

JAN

2.700

0.008

2.157

0.000

2.514

FEB

4.110

0.43G

2.270

1.380

2.830

MAR

1.920

0.008

1.618

0.10G

2.544

APR

2. GOO

0.179

1.991

0.7GS

2.735

MAY

5.420

0.G44

3.374

0.971

2.742

JUN

5.290

1.254

3.153

1.257

2.351

JUL

4.150

0.58G

4.72G

0.000

0.901

AUG

5.780

0.192

G.440

0.000

0.9G9

SEP

1.850

0.000

2.199

0.000

0 . 457

OCT

0.3S0

0.000

0.554

0.000

0.035

NOU

1.160

0.000

0.782

0.000

0.282

DEC

4.920

0.209

l.GGG

0.433

1.484

TOT 40.2G0 3.51G 30.929 4.916 1.G59

Figure 11-13.— Sample output of monthly totals of rainfall,
runoff, evapotranspiration, percolation, and average soil
water, and averages for the period of simulation.

189

MONTH

RAIN

1975

RUNOFF

PERC

AUG SW

JAN

5.020

0.490

2.321

2.522

2.977

FEB

7.170

1.035

2.4G0

3. Gil

2.S20

MAR

9.780

2.9G1

2.930

3.G82

2.SG4

APR

3.930

0.934

2.173

1.2G0

2.717

MAY

G.070

0.G33

3.123

1.G8G

2.771

JUN

3.550

1.071

2.93G

1.0G7

2.479

JUL

4.G70

0.043

G.012

0.000

0.G15

AUG

2.340

0.000

2.194

0.000

0.1G5

SEP

5.370

0.339

2.934

0.000

1.200

OCT

0.350

0.000

0.S1G

0.000

2.157

NOU

0.000

0.329

0.000

1.721

DEC

0.000

0.2G2

0.000

1.431

TOT

48.250

7.505

23.G02

13.823

2.001

ANNUAL AUERAGES

MONTH

TOT

RAIN

44.255

RUNOFF

5.511

29,755

PERC

9.372

AUG SU

JAN

3.8G0

0.249

2.239

1.2G1

2.745

FEB

5.G40

0.735

2.3S5

2.49G

2.875

MAR

5.850

1.485

2.274

1.894

2.704

APR

3.2G5

0.55G

2.082

1.014

2.72G

MAY

5.745

0.G39

3.249

1.329

2.756

JUN

4.420

1.1G3

3.074

1.162

2.415

JUL

4.410

0.314

5.3G9

0.000

0.758

AUG

4.0G0

0.09G

4.31T

0.000

0.5G7

SEP

3.G10

0.170

2.532

0.000

0.828

OCT

0.355

0.000

0.G85

0.000

1.126

NOU

0.580

0.000

0.55G

0.000

1.001

DEC

2.4S0

0.105

0.9G4

0.217

1.458

1.830

Figure 11-13. — Sample output of monthly totals of rainfall, runoff,
evapotranspiration, percolation, and average soil water, and
averages for the period of simulation--Conti nued.

APPLICATION FOR EVALUATION OF MANAGEMENT OPTIONS

In hydrology option 2, changes in the parameters for infiltration and for
surface flow can reflect rather directly the changes resulting from management
options. Table 11-10 indicates some general effects of management on infiltra-
tion parameter RC. A denser canopy or a denser stem count (grass in comparison
to a row crop) generally shows a larger RC. The fallow season RC values are
estimated as 0.8 of that for the growing season RC. No-till practices with
good mulch cover should make year-round values for RC relatively stable. If
mulch cover is poor for no-till practices, the RC value should be estimated
lower because the soil surface is expected to crust.

190

Table 11-10. — Effect of cultural practices on model parameters

p~~T- Effective Effective ,, Effective Conductivity,

practices slope roughness, RMN1' length, XLP RC

Grassed concentra- c,m^ - a_ c,m c„

. , j-n . t Same Increase Same Same.

ted flow channel .

Standard terraces Decrease Same Increase Same.

Chisel-tillage -Same Large increase Same Increase.

N^11 Same Increase Sarae °™hat.

Contour plow Decrease Same Increase Same.

1/ High roughness implies low value of Cc

Cultural practices that affect routing of the runoff water on the field
surface also significantly affect runoff. Contour plowing, for example, can
extend considerably the effective path of overland flow, XLP. This reduces the
peak outflow, which dramatically affects the amount of erosion.

In evaluating the effective flow length, XLP, the actual mean overland
flow path should be measured from a map, when available. A fraction of the
mean channelized flow length also should be added. The fraction used should be
small for steeper, smoother channels, and large (near 1.0) for rough (grassed
or vegetated) and flat sloped channels. Terraces effectively increase storage
volume along the flow path, which can be simulated for calculating runoff in
this model by increasing length of the flow path and decreasing the effective
slope.

Management practices affect both hydrology options through changes in leaf
area index. More intense management (such as high fertilization rates) in-
creases crop production and LAI. Increasing LAI causes greater water use which
reduces soil water storage. Runoff is reduced when soil water storage is low-
ered. The LAI values in table 1 1-8 are for high management levels that produce
large plants, and they should be reduced for less intense management. The LAI
values in table 1 1-8 generally should be multiplied by 0.83, 0.67, and 0.5 for
good, fair, and poor management, respectively.

Hydrology option 1 also reflects management changes through change of the
SCS curve number for AMC-II condition. Tabulated values of CN2 are given as a
function of soils, land use, and management level in the SCS Hydrology Handbook
( 5_) . Recent work on the effects of residue and tillage on the SCS curve num-
ber (_3) provides more refined estimates of CN2 for such modern management prac-
tices as conservation tillage and no-till systems.

191

REFERENCES

(1) Holtan, H. N., C. B. England, G. P. Lawless, and G. A. Schumaker.

1968. Moisture-tension data for selected soils on experimental water-
sheds. U.S. Department of Agriculture, Agricultural Research Service,
ARS 41-144. 609 pp. (Series discontinued; Agricultural Research
Service now Science and Education Administration-Agricultural
Research. )

(2) Kothandaraman, V., and R. L. Evans.

1972. Use of air-water relationships for predicting water temperature.
Illinois State Water Survey, Report of Investigations No. 69, Urbana,
111. 14 pp.

(3) Rawls, W. J., C. A. Onstad, and H. H. Richardson.

1979. Residue and tillage effects on SCS runoff curve numbers. (Accep-
ted for publication in the Transactions of American Society of Agri-
cultural Engineers.)

(4) U.S. Department of Agriculture.

1968. Hydrologic data from experimental agricultural watersheds. Mis-
cellaneous Publication No. 1330.

(5) , Soil Conservation Service.

1972. National Engineering Handbook. Section 4. Hydrology. 548 op.

(6) U.S. Department of Commerce.

1968. Climatic atlas of the United States.

(7)
(8)
(9)

1979. Climatological data, State of (Arizona).

1979. National summary, climatic data.

1979. Hourly rainfall data, State of (Arizona).

192

Chapter 2. A MODEL TO ESTIMATE SEDIMENT YIELD FROM FIELD-SIZED AREAS:
SELECTION OF PARAMETER VALUES

G. R. Foster, L. J. Lane, and J. D. Nowlin-/

INTRODUCTION

The erosion/sediment yield component of CREAMS discussed in this chapter
is for use by planners and managers who select practices to control nonpoint
pollution due to sediment coming from field-sized agricultural areas. This
model combines new modeling concepts with such commonly accepted relationships
as the Universal Soil Loss Equation (USLE) to provide a flexible, powerful
model requiring a reasonable number of inputs. The model computes erosion,
sediment yield, and particle composition of the sediment on a storm-by-storm
basis. Long-term effects are evaluated by simulating over a long record. Main
inputs are rainfall erosivity and runoff for each storm and erosion-sediment
transport characteristics of the area. Effects of spatial variability in a
downslope direction can be analyzed.

The model is based on the fundamental concept that if sediment available
from detachment is less than transport capacity, detachment controls sediment
yield. Conversely, if sediment load exceeds transport capacity, transport ca-
pacity controls sediment yield.

MODEL STRUCTURE

The model is structured around three basic elements: overland flow; con-
centrated (channel) flow; and an impoundment (pond). The study area is repre-
sented by a sequence of these elements. The overland flow element is called
first, followed by a channel or pond element, or both, if these additional
elements are required.

IMPLEMENTATION OF THE MODEL

The model is programed in standard FORTRAN. The main program, which is a
control program, calls subprograms that read data, calculate erosion and sedi-
ment yield, and display the output.

1/The authors are, respectively, hydraulic engineer, USDA-SEA, Agricul-
tural Engineering Department, Purdue University, West Lafayette, Ind.; hydro-
logist, USDA-SCA, Southwest Rangeland Research Watershed Center, Tucson, Ariz.;
and computer programer, Agricultural Engineering Department, Purdue University,
West Lafayette, Ind.

193

PROGRAM FLOW

The program operates over a series of storms but takes each storm individ-
ually. The program uses two data input files. One file contains the hydrolog-
ic input data, that is, storm erosivity (EI), volume of runoff, and peak runoff
rate. The second input file contains inputs that characterize the erosion and
sediment transport characteristics of the area (for example, soil erodibility,
hydraulic roughness, slope shape). Sediment yield is computed by the program
calling the elements in the sequence defined by the user. Output is sediment
load and concentration of each particle type.

Factors that change with time are updated periodically. The date for each
storm is compared with the date entered for the erosion parameter values. If
the date for the storm exceeds the last date that current parameter values
apply, new values are read for parameters that change. If updating is unneces-
sary, the program proceeds with computations for the next storm by reusing
values from the previous storm.

SUBPROGRAMS
Main

The main subprogram is actually the control section of the overall pro-
gram. It calls subroutines for input, output, and erosion and transport compu-
tations for the elements.

Input

Subroutines read data from the input files and convert all input data to
units of feet, seconds, and pounds. This reduces confusion in using common
variables among different subroutines. The unit for length variables is feet
anywhere in the program, for example, except for input or output where the
variables are in customary units for the user's convenience.

Output

Subprograms print out detailed results, as requested by the user.

Overland Flow

The overland flow subprogram computes interrill-ril 1 (sheet-rill) erosion
and sediment transport by overland flow. A modified version of the USLE has
separate terms for detachment caused by flow and detachment caused by the im-
pact of rain drops. The relationship uses input values for storm EI, volume of
runoff, and peak discharge rate for the storm, and it uses USLE factors for
soil erodibility, cover-management, and contouring.

The Yalin sediment transport equation is used to compute transport capaci-
ty. Rate of deposition is assumed to be directly proportional to the differ-
ence between transport capacity and sediment load. The model uses size and
density of particles to estimate selective deposition. Hydraulic roughness for
the overland flow surface characterizes the effect of roughness and vegetation
on transport capacity.

194

The subprogram calls the subroutine PROFILE, which constructs a concave,
convex, or a complex slope from slopes at the beginning, midsection, and end of
the hi 1 1 si ope profile supplied as input. The program defines three slope seg-
ments for any convex slope shape, 10 segments for any concave slope shape, and
a single segment for any uniform section.

The subprogram merges coordinates for these segments, along with coordi-
nates where soil erodibility, cover-management, contouring, and hydraulic
roughness change, into a single array of coordinates. Even when a uniform
slope is specified, the user can consider changes in soil erodibility and the
other factors along a slope. Computations proceed downslope segment by seg-
ment. The amount of sediment produced by detachment from intern' 11 erosion is
calculated and added to that arriving from upslope segments. The sum (poten-
tial sediment load) is compared with transport capacity. If transport capacity
exceeds the potential sediment load, no deposition occurs and detachment by
flow occurs at a capacity rate or a rate that will just fill transport capa-
city. If transport capacity is less than the potential sediment load, however,
deposition occurs.

Sediment composed of up to five particle types can be considered. For
most soils, particles are eroded as both aggregates and primary particles.
Primary particles are sand, silt, and clay, and aggregates are conglomerates of
primary particles and organic matter. A high percentage of sediment for silt
loam soils is aggregated. The user supplies information on particles (density
and diameter), or the model will estimate a distribution from the distribution
of primary particles of the soil mass.

Principal output from the overland flow component is sediment load and
concentration for each particle type for the storm. These values are final if
overland flow is the only element called in the sequence. Otherwise, overland
flow output is input for downstream elements.

Concentrated (Channel) Flow

The channel subprogram represents detachment and sediment transport in
terrace channels; waterways; and small intermittent streams. Flow concentra-
tions also include areas through the middle of a field where overland flow con-
centrates due to natural topography. Flow also may concentrate along field
boundaries where a ridge on the outside of the field causes overland flow to
collect along the edge of the field. Grass or a ridge at the field outlet also
may slow the flow, causing deposition in the backwater.

The initial section of this subprogram sets up increments along the chan-
nel equal to 0.1 of the channel's effective length, the length of the channel
if it were long enough to have zero flow at its upper end with the assumed
lateral inflow. Some channels begin with an initial flow rate where overland
flow area is above the entrance to the channel. Additional increments are de-
fined if changes, such as in cover, occur along the channel.

The program selects from a variety of dimensionless curves to compute
friction slope. This selection is based on channel slope and outlet control.
If outlet control causes backwater, one curve from a group approximates the
decrease in slope of the energy gradeline. If critical flow controls at the
outlet, the program selects one curve from three applicable curves. Three

195

curves pertain to a channel having zero slope, but friction slope may be as-
sumed to equal channel slope.

The channel is assumed to be triangular (user supplies sideslope) or rec-
tangular (user supplies channel width). A third option is for the program to
compute eroded widths for a rectangular channel.

Computations proceed downslope as in the overland flow subprogram. Exact-
ly the same concept of detachment or transport limiting is used to route the
sediment downstream.

Pond

The pond subprogram estimates deposition of sediment in impoundment ter-
races having controlled pipe outlets. Deposition in shallow natural impound-
ments caused by a ridge around a field, heavy vegetation at the outlet to the
field, or a pipe culvert is analyzed with the channel element and a backwater
curve. The deposition relationship for the pond element is basically an expo-
nential decay function with parameter values related to volume of runoff,
geometrical characteristics of the impounded area, discharge rate from the im-
pounded area, infiltration rate over the pond area, and size and density dens-
ity of sediment particles.

MAJOR ASSUMPTIONS

This model, like any other model, is based on many assumptions. The user
should be aware of the most significant assumptions because in some
applications the model is invalid.

Profile

The curved portions of a land profile are assumed to be described by a
quadratic equation where the end slopes are the same as the adjoining uniform
slopes. The actual field slope may not be duplicated, but the essential
effects of concavity, convexity, and complexity are included.

Di scharge

Discharge at any point in the watershed is assumed to be directly propor-
tional to the drainage area above that point. Overland flow discharge at a
location is computed, therefore, as a product of length of slope to that point
and maximum excess rainfall rate which is attenuated for nonuniform rainfall
rates and travel time. This attenuated peak discharge is used as a
characteristic discharge for the runoff event.

Erosion and Transport on Overland Flow Areas

The relationship to estimate detachment is on a storm basis, while trans-
port is estimated on an instantaneous discharge basis. Sediment concentration
in the flow is assumed to be the average concentration for the storm. Concen-
tration for detachment is determined by dividing the amount of sediment de-
tached for the storm for a segment by total amount of runoff per unit area.
The characteristic discharge multiplied by the concentration gives rate of soil
loss (per unit time) at the characteristic discharge. Transport also is compu-
ted with the characteristic discharge rate so that detachment and transport
will be on the same basis.

196

An assumption is necessary to deal with simultaneous deposition and de-
tachment. Whether the flow is detaching or depositing, the model always as-
sumes that intern' 11 erosion adds sediment to the flow. On a given segment,
the potential sediment load is computed by adding detachment from intern" 11
erosion to the incoming sediment load from the next upslope segment. If this
potential sediment load exceeds transport capacity on the segment, deposition
occurs on the segment. If deposition occurs, no rill erosion is allowed. If
transport capacity exceeds this sediment load, two other possibilities exist.
The first is that rill erosion can occur at its capacity rate and still give a
total sediment load less than the transport capacity. The second possibility
is that if rill erosion were to occur at its capacity rate, total sediment load
at the end of the segment would exceed the transport capacity. In this situa-
tion, rill erosion is limited to that which would just fill transport capacity.
This concept is more realistic than assuming that rill erosion occurs at its
capacity rate even when deposition occurs. To allow simultaneous rill erosion
and deposition at the same time is a conceptual inconsistency for erosion and
transport over cohesive agricultural soils.

The capacity of overland flow to transport sediment is estimated using the
Yalin bedload sediment transport equation. If desired, the user can increase a
constant in the equation to account for both bedload and suspended load. Of
several sediment transport equations considered, the Yalin equation appeared to
be as good or better than most for transport by overland flow, especially when
small particles and particles having specific gravities less than that of sand
are considered.

The Yalin equation is modified to consider particle mixtures. If the sedi-
ment load of each particle type exceeds the transport capacity of the respec-
tive particle type, sediment transport capacity is distributed among the par-
ticle types based on transportability of the particles. If sediment load of a
particular type is less than the transport capacity for that type, its excess
transport capacity is shifted to particles having a deficit transport capacity.
This modification prevents a small load of a particular particle type from
having more than its share of the total transport capacity. When deposition
occurs, rate of deposition is assumed to be directly proportional to the dif-
ference between transport capacity and sediment load. The proportionality con-
stant is assumed to be directly proportional to fall velocity of the particle
divided by the product of flow velocity and flow depth. This gives an exponen-
tial decay for rate of deposition as a function of distance.

Any deposited particles are assumed to become reattached immediately to
the soil mass, that is, deposited particles are unavailable as detached par-
ticles for subsequent transport without being redetached. Likewise, tillage is
not assumed to produce a supply of detached particles that is depleted over
time by transport. Increased erosion from tillage is analyzed by adjusting the
USLE soil loss ratio.

Erosion and Transport in Channel Flow

Input to the channel is a uniform lateral inflow of runoff and sediment
from an overland flow or another channel element. The characteristic discharge

197

is used to compute detachment, sediment transport, deposition, and sediment
concentration in the channel elements.

Outlet conditions for the channel are assumed to be controlled by a
downstream uniform flow, critical depth, or a structure having a known rating
curve (for example, a flume restriction in an experimental watershed or a field
boundary). Subcritical flow is assumed unless the option is specified that
slope of the energy gradeline (friction slope) equals the channel slope.

Since many channels in farm fields may be approximated as triangular chan-
nels, a triangular channel with 5:1 sideslopes was used to develop the friction
slope curves. Therefore, the actual channel must be approximated by a triangu-
lar channel to compute the friction slope. Remaining channel computations are
made assuming a triangular, rectangular, or eroding channel section. The tri-
angular and rectangular channel sections may have cover, but the eroding chan-
nel section is assumed to be bare with no cover. Enlargement occurs in the
eroded channel and section properties remain fixed in the triangular channel.
Width of the rectangular channel increases once the computed eroded width
exceeds the initial width read as input.

Concepts for detachment and transport in the channel are exactly the same
as those for overland flow. Lateral inflow of sediment in the channels is
equivalent to interrill erosion, and channel erosion is equivalent to rill
erosion. Relationships for the detachment capacity of channel erosion are com-
puted using expressions developed from an experimental and analytical rill ero-
sion study by Lane and Foster (vol. Ill, ch. 11). The algorithm considers the
influence of a nonerodible boundary at some depth below the bottom of the
channel. When a channel erodes to the nonerodible boundary, the channel widens
and erosion rate decreases with time. This frequently occurs in many
midwestern fields at planting time in areas where grass waterways should be
installed.

The effect of tillage on erosion by channel flow is modeled by assuming
that tillage greatly decreases the critical shear stress for detachment to be-
gin. Tillage is assumed to loosen, (that is, decrease the critical shear
stress) down to a given depth. No erosion is allowed below this depth. Criti-
cal shear stress increases after tillage as the soil consolidates from traffic,
wetting and drying, and other processes. Values for critical shear stress and
soil erodibility by channel flow have not been validated substantially.

Channel erosion does not occur throughout the duration of a storm. De-
tachment occurs only when shear stress exceeds critical shear stress. This
time is estimated by assuming that the shear stress is linearly distributed in
time. Detachment is assumed to occur at a rate based on the characteristic
discharge for the period that shear stress exceeds the critical shear stress.

Transport Through Impoundments

These relationships for the impoundment component were derived from a de-
tailed model (3.) based on settling theory in still water. Output from the de-
tailed model fit observed experimental data well. Regression analyses were
used to fit the relationships used in this model to output from the detailed
model .

198

This component applies to impoundment terraces that drain completely after
a runoff event and to other small impoundments where discharge is controlled by
an orifice restriction in an outlet pipe. Although several other impoundments
occur behind ridges around fields, pipe culverts, and farm ponds, the impound-
ment component generally should not be applied to these situations. The chan-
nel element with backwater is recommended for impoundments by ridges and cul-
verts. Relationships for standard reservoir deposition may be applied to farm
ponds, using the model output to estimate runoff reaching the pond.

MODEL INPUTS AND PARAMETERS

The model inputs are the hydrologic variables rainfall storm erosivity
(EI), volume of runoff, and characteristic peak excess rainfall rate (peak run-
off rate at the outlet divided by area). These generally are obtained from
the hydrology component of CREAMS or from observed data. Table 11-11 shows the
hydrology pass file variables, format, and sample data for the input to the
erosion and sediment yield program. Figure 11-14 a represents card image
format for the pass file from the hydrology model. The model parameters
characterize the erosion-sediment transport-deposition features of the area.
These values come from a variety of sources. Tables 11-12 and 11-13 identify
inputs and parameters, possible sources, and indications of the quality of a
parameter estimate.

PREPARATION OF INPUT DATA FILES

This section shows how to assemble input data files by briefly describing
the parameters and their location in the data set. For a more comprehensive
definition of the parameters and a method of selecting values, refer to the
following section of this publication.

The model reads input from two separate files: a hydrology file and a pa-
rameter file. Unless specifically stated otherwise in the card description,
all numeric input formats consist of 8-column fields. Integers are read with
18 formats, and real numbers are read with F8.0 formats. Integers must be
right justified in columns 1-8, 9-16, 17-24, ..., 73-80. Real numbers must be
contained within these same columns, and the decimal point must be entered in
the number. A sample value is given after each parameter is defined. If the
sample has a decimal, the parameter is real; otherwise, it is an integer. The
alphanumeric input is read with A4 formats. Specific instructions are given
whenever alphanumeric input is required. A schematic representation of the
parameter data deck is shown in figure 11-15. Figure 11-16 is a schematic data
deck with sample data for 3 years on watershed P-2 at Watkinsville, Ga.

Blank cards or blank entries are used on some data cards to indicate a zero
entry. If your computer does not read blanks as zeroes, enter zeroes instead
of leaving those parameters blank. Some computers can be set using control
statements to read blanks as zeroes.

If some updateable parameter does not change when others change, the last

value read in for the parameter will be used by the program, if desired. When

this option is used, data cards for these parameters are omitted. This is dis-
cussed in greater detail in the sections on updateable parameters.

199

Table 11-11 — Hydrology pass file description and data for input to the
erosion/sediment yield model

A. Storm/Hydrology Data File

Card 1. SDATE, RNFALL, RUNOFF, EXRAIN, EI, DP, PERCOL, AVGTMP, AVGSVvC,

ACCPEV, POTPEV, ACCSEV, POISEV

SDATE

RNFALL

RUNOFF

EXRAIN

PERCOL
AVGTMP

AVGSVvC
ACCPEV

POTPEV

ACCSEV

POTSEV

Date of storm (Julian date) , e.g. 73148

Volume rainfall (in), e.g. 4.27

Volume of runoff (in), e.g. 1.58

Characteristic excess rainfall rate (in/hr), e.g. 4.13

Wischmeier English EI for the given storm, e.g. 67.41

Number of days since the last storm when percolation
occured, e.g. 1

Percolation below the root zone (in), e.g. 1.015

Average temperature between storms (Degrees F.), e.g.
72.8

Average soil water between storms (in/in), e.g. 0.3239

Actual EP (evaporation from plants) for the period
between storms (in), e.g. 0.02210.056

Potential EP for the period between storms (in)
0.02210.056

e.g.

Actual ES (evaporation from soil) for the period
between storms (in), e.g. 0.000|0.000

Potential ES for the period between storms (in), e.g.
0. 00010.000

Card 1 is repeated for each rainfall event. The last card in the file
should be blank to indicate the end of data. The Hydrology program creates a
file called "hYDPAS" specifically for use as this file. The values in the
Storm/ hydrology file from DP to POTSEV are only read into the Erosion program
so they can be passed through to the Chemicals program. If a file for the
Chemicals program isn't going to be created then the only values required for
the Storm/ hydrology file are SDATE, RNFALL, RUNOFF, EXRAIN, and EI.

200

Table 11-11. — Hydrology pass file description and data for input to the
erosion/sediment yield model --continued

A small sample of a typical Storm/ hydro logy Data file follows to illus-
trate the file structure.

Bormat(l6,F6.2,F6.2,F6.2,F6.2,I2,F6.2,F6.2,F6.4,F6.3,F6.3,F6.3,F6.3)

73148

4.2700

1 8686

37587

.9727 1

770

1221

3378

1 .3266

.9891

73156

.2800

7860 1

.002

1801

.3577

1 .9595

.2168

73157

1 . 2200

. 1583

. 4580

5645 1

.585

8851

.3672

2 0836

4875

75159

. 6000

0128

.0540

.5388 2

.302

0943

3653

2.3687

0597

201

Table 11-12. Model input

Variable Source

Runoff Volume (V)--------------- Estimated by a model.

Characteristic runoff rate (^p) -------- Estimated by a model.

Storm erosivity (EI)-------------- Estimated from volume of rain-
fall and maximum 30-min in-
tensity or volume of rainfall
alone.

(BLANK CARD FLAGS THE END OF THE FILE)

REMAINDER OF THE

STORM/HYDROLOGY DATA

(I CARD/EVENT)

FORMAT(16,F6,2,F6,2,F6,2,F6,2,I2,F6,2,F6,2,F6,4,F6,3,F6,3,F6,3,F6,3)

SDATE RNFALL RUNOFF EXRAIN El DP PERCOL AVGTMP AVGSWC ACCPEV POTPEV ACCSEV POTSEV

Figure 11-14. — Sample format and card image arrangement for hydrology pass
file, from either hydrology model option.

Some cards, such as card 13, contain unnecessary information. If a rating
curve is specified, for example, the first four parameters on card 13 are un-
necessary. These may be left blank, assigned zero, or assigned an obviously
incorrect value, such as 999. This latter entry could help trace input er-
rors.

Since all data files must be input in English units, values on the sample
card decks are shown with English units. The model is written for variables
with English units.

SELECTION OF INPUT VALUES

Input values generally can be selected from readily available information
from such sources as a soil survey, topographic maps, aerial photographs, soil
description, cropping history, and a site visit. Note the following input re-
quirements and assemble the required source materials.

Obviously, the model is more sensitive to some parameters, as discussed in
volume I, chapter 6. The sensitive variables require more careful selection.
If sediment yield is primarilly controlled by detachment, overall the detach-
ment parameters are more important whereas transport parameters are more impor-
tant when deposition primarily controls sediment yeild. However, for specific
locations, detachment may limit sediment yield for some storms, while transport
will limit for other storms. Detachment may limit on one part of the watershed
while transport will limit on another part. Detachment may control sediment
yield for the fines while at the same time transport will control the yield of
coarse particles. The result is mixed control between detachment and transport
parameters, preventing general statements on the sensitivity of particular par-

202

Table 11-13. — Erosion model parameters
Parameter Definition

v ------ Kinematic viscosity

n. - - - — Manning's n for overland
flow over bare smooth
soil (fine seedbed) .

n. . ----- Manning's n for channel
flow over bare, smooth
soil (fine seedbed).

p ., - - - - Weight density of soil
mass.

K . Soil erodibility factor

for channel erosion.

C , — - — Constant in Yalin sedi-
ya ment transport equation.

Sand, silt,- - Primary particle distri-
clay. bution of original soil

mass.

Particle Particle size class and

character- density of particle,

istics.

X.„ Overland flow slope

ov length.

S ______ Average overland flow

slope steepness.

Sb ----- Slope at beginning of

overland flow profile.

S ----- Slope at middle of over-
land flow profile.

S ----- slope at end of overland
flow profile.

x3' y3 — - - Coordinates of mid-

x-, y. uniform slope section.

A ----- Overland flow area

K ------ soil erodibility factor

(rill-interrill erosion)

C __--__ Cover-management factor

(rill-interrill erosion)

P ______ Contouring factor

(rill-interrill erosion)

n — - - - Manning's n for overland
flow over a covered soil
surface.

definitions, and sources and quality of estimates

Source of estimate

Quality of Estimate

Excellent. However, only parame-
ter expressing temperature
effect. Quality for expressing
that effect unknown.

Good but subjective.

Poor. May require calibration,

Good. Supposedly fixed, but may
require calibration.

Very good.

Good for most midwestern silt loam
soils; unknown for most other
soils.

Handbook

Model manual

Soil survey and
experience.

Model manual

Soil survey, soil
tests experi-
ence.

Model manual and
soil survey in-
formation or cal-
culated from model
equations using
primary clay, silt
and sand.

Maps, soil survey, Good, but problem of choosing re-
field observation. presentative length.

Maps, soil survey,
field observation.

Maps, soil survey
field observation.

Maps, soil survey,
field observation.

Map

Model manual ; also
USLE Handbook.

Model manual

Good, but problem of choosing re-
presentative length.

Good, but problem of choosing re-
presentative steepness.

Good, but problem of choosing re-
presentative section.

Very good.

Good, based on extensive plot
data.

Poor; value poorly defined for in-
dividual storms.

Good, but subjective.

203

Table 11-13.

■rosion model parameters, definitions, and sources and quality of estimates-
continued

Parameter Definition

Shape Channel shape

\ . ----- -Channel length

A . Drainage area draining

p into upper end of chan-

nel .

A . , - - - - -Area drained by channel

Outlet - - -Outlet control parame-
control. ters including channel
width, sideslope, lon-
gitudinal slope,
Manning's n, rating
curve.

Slope - - - - -Slope along channel

n . ----- -Manning's n for channel

with cover.

t - — — -Critical shear stress

which erosion begins in
channel .

r - - - - -Critical shear stress
for cover breakup.

d ----- -Depth to nonerodible

layer in channel .

d . , - - - - -Depth to nonerodible

layer at side of chan-
nel .

w . ----- -Channel width

I _____ _ -Channel sideslope

F , B - - - - -Coefficients for pond
surface area vs depth
intake rate.

i _____ _ -intake rate

d ----- -Diameter of orifice

in outlet pipe.

A , ----- -Drainage area above

pond.

Source of estimate

Quality of Estimate

Experience and
field observation.

Map, field obser-
vation.

Map

Experience, field
observation,
model manual .

Map, field obser-
vation.

Model manual , hand-
books provided;
nbch selected from
same handbook.

Model manual , ex-
perience.

Model manual , ex-

8BOT_!to^eld

Model manual , ex-
perience, field
observation.

Model manual ,

field observa-
tion photo.

Model manual ,
field observa-
tion map.

Field survey mod-
el manual , map

Soil survey, ex-
perience.

Design notes,
field observa-
tion experience.

Map

Good, but subjective.

Good, but can be quite subjective.

Very good.

Poor and highly subjective.

Very good.

Good, but subjective.

Poor, values not known for many
agricultural soils and manage-
ment effects not known.

Fair for nonincorporated residue,
poor for incorporated.

Fair, but subjective.

Poor and highly subjective.

Fair to good.

Excellent with field survey, good
with other means of estimating.

Excellent or good if based on ex-
perience.

Excellent.

204

205

206

ameters. Generally, a sensitivity analysis is advisable for each specific
problem.

Discussion of selection of input values generally parallels the layout of
the input data cards. Some input variables are not discussed because selection
of a value is obvious.

Storm Hydrology Input

Hydrology Pass File

SPATE— The Julian calendar in table 11-14 may be used to convert an ordinary
calendar date. This date is given by specifying the last two digits of the
year (for example, 78) followed by the Julian date (for example, 094 for April
4, or 78094).

RNFALL— Rainfall volumes for a series of storms are available from rainfall

records of the National Weather Service and other agencies that collect weather
data in your locale. Breakpoint data are most desirable, although hourly or
daily data may be used.

RUN0FF--This is runoff volume per unit watershed area for the storm. Runoff is
assumed to be uniform over the drainage area.

EXRAIN— The characteristic peak excess rainfall rate for the storm is obtained
by dividing peak discharge at the watershed outlet by watershed area. If it is
computed by subtracting infiltration rate from rainfall rate, it must be atten-
uated to account for nonuniform rainfall rates and time of travel. In the ero-
sion/sediment yield component of CREAMS, characteristic peak runoff rate at any
point in the watershed is taken as directly proportional to the drainage area
above that point.

J£I_--The EI variable, as defined by Wischmeier and Smith (12), is a measure of
rainfall erosivity. If a rainfall hyetograph of the given storm is available,
EI for the storm may be estimated from the following procedure. Divide the
rainfall hyetograph into periods so that rainfall intensity may be assumed to
be constant for a period. For each period, calculate the unit rainfall energy
per unit of rainfall from

e = 916 + 331 logioi ClI-5]

where e = unit rainfall energy (ft-tons/acre-in of rain) and i = rainfall in-
tensity (in/hr'
to obtain the energy

The storm energy

100 gives EI in

Wischmeier' s English EI units.

i 1 1 i a i i cue i <j_y \ i i-i,uiii/ av,i c" in ui i a i 1 1 ; anu i - rainiaii 111-

Multiply the unit energy by the rainfall amount in the period
,.v.rgy for that period. Add these incremental energies for all
periods to obtain the total rainfall energy for the storm,
multiplied by the storm's maximum 30-min intensity divided by

i u i • _ „ I _ i- i • _ i_ r-T £ j. _

Where the rainfall hyetograph is unavailable, total storm energy, E, may
be estimated by computing e (unit rainfall energy) using the maximum 30-min
storm intensity and multiplying by volume of rainfall. This value multiplied
by maximum 30-min intensity is an estimate of EI. If the maximum 30-min inten-
sity is not known but the maximum 60-min intensity is available, multiply the
maximum 60-min intensity by 1.6 to estimate the maximum 30-min intensity.

207

These detailed data may be unavailable for a given storm. The best avail-
able information may be hourly or daily rainfall amounts, which are, by them-
selves, poor estimates of rainfall erosivity (6J . A good estimate of intensity

poor ebLimaueb ui raiiuaii eiubivity v u^ . n yuuu eb t imaut: ui mierib i lj

red for good erosion estimates. Given only rainfall amount, however

mas/ ho octi'matoH fynm-

is requi. _

storm EI may be estimated from

EI = 8.00 V 1-51
r

[1 1-6]

where V = volume of rainfall (in). Since the coefficient of determination
r

(R^) for this equation is 0.54, EI values from the equation are subject to con-
siderable error for any given specific storm.

DP, PERCOL, AVGTMP, AVGSWC, ACCPEV, POTPEV, ACCSEV, and POTSEV— These fields
are not used by the erosion/sediment yield component. They may be left blank
unless the erosion program is used to construct an input data file for the
chemical program.

Table 11-14 — Julian Calendar^

* * * *

* January

*****

1/ 1

2/ 2

3/ 3

4/ 4

5/ 5

6/ 6

7/ 7

8/ 8

9/ 9

10/10

11/11

12/12

13/13

14/14

15/15

16/16

17/17

18/18

19/19

20/20

21/21

22/22

23/23

24/24

25/25

26/26

27/27

28/28

29/29

30/30

31/31
* * * *

* February

*****

32/ 1

33/ 2

34/ 3

35/ 4

36/ 5

37/ 6

38/ 7

39/ 8

40/ 9

41/10

42/11

43/12

44/13

45/14

46/15

47/16

48/17

49/18

50/19

51/20

52/21

53/22

54/23

55/24

56/25

57/26

58/27

59/28

For

leap years,

add 1 day
* * *

to Jul ian
* * March i

date folic

r * * * *

)wing Feb

ruary 28.

60/ 1

61/ 2

62/ 3

63/ 4

64/ 5

65/ 6

66/ 7

67/ 8

68/ 9

69/10

70/11

71/12

72/13

73/14

74/15

75/16

76/17

77/18

78/19

79/20

80/21

81/22

82/23

83/24

84/25

85/26

86/27

87/28

88/29

89/30

90/31
* * * •

k * April "

r * * * *

91/ 1

92/ 2

93/ 3

94/ 4

95/ 5

96/ 6

97/ 7

98/ 8

99/ 9

100/10

101/11

102/12

103/13

104/14

105/15

106/16

107/17

108/18

109/19

110/20

111/21

112/22

113/23

114/24

115/25

116/26

117/27

118/28

119/29

120/30

* * *

* * May *

* * * *

121/ 1

122/ 2

123/ 3

124/ 4

125/ 5

126/ 6

127/ 7

128/ 8

129/ 9

130/10

131/11

132/12

133/13

134/14

135/15

136/16

137/17

138/18

139/19

140/20

141/21

142/22

143/23

144/24

145/25

146/26

147/27

148/28

149/29

150/30

151/31

208

Table 11-14.— Jul ian calendar—continued

* * *

* * June *

* * * *

152/ 1

153/ 2

154/ 3

155/ 4

156/ 5

157/ 6

158/ 7

159/ 8

160/ 9

161/10

162/11

163/12

164/13

165/14

166/15

167/16

168/17

169/18

170/19

171/20

172/21

173/22

174/23

175/24

176/25

177/26

178/27

179/28

180/29

181/30

* * *

* * July *

* * * *

182/ 1

183/ 2

184/ 3

185/ 4

186/ 5

187/ 6

188/ 7

189/ 8

190/ 9

191/10

192/11

193/12

194/13

195/14

196/15

197/16

198/17

199/18

200/19

201/20

202/21

203/22

204/23

205/24

206/25

207/26

208/27

209/28

210/29

211/30

212/31

* * * i

* * August *

r * * * *

213/ 1

214/ 2

215/ 3

216/ 4

217/ 5

218/ 6

219/ 7

220/ 8

221/ 9

222/10

223/11

224/12

225/13

226/14

227/15

228/16

229/17

230/18

231/19

232/20

233/21

234/22

235/23

236/24

237/25

238/26

239/27

240/28

241/29

242/30

243/31
* * * *

* September

< * * * * i

244/ 1

245/ 2

246/ 3

247/ 4

248/ 5

249/ 6

250/ 7

251/ 8

252/ 9

253/10

254/11

255/12

256/13

257/14

258/15

259/16

260/17

261/18

262/19

263/20

264/21

265/22

266/23

267/24

268/25

269/26

270/27

271/28

272/29

273/30

* * * i

* * October

*****

274/ 1

275/ 2

276/ 3

277/ 4

278/ 5

279/ 6

280/ 7

281/ 8

282/ 9

283/10

284/11

285/12

286/13

287/14

288/15

289/16

290/17

291/18

292/19

293/20

294/21

295/22

296/23

297/24

298/25

299/26

300/27

301/28

302/29

303/30

304/31
* * * *

* November

*****

305/ 1

306/ 2

307/ 3

308/ 4

309/ 5

310/ 6

311/ 7

312/ 8

313/ 9

314/10

315/11

316/12

317/13

318/14

319/15

320/16

321/17

322/18

323/19

324/20

325/21

326/22

327/23

328/24

329/25

330/26

331/27

332/28

333/29

334/30

* * * *

* December

*****

335/ 1

336/ 2

337/ 3

338/ 4

339/ 5

340/ 6

341/ 7

342/ 8

343/ 9

344/10

345/11

346/12

347/13

348/14

349/15

350/16

351/17

352/18

353/19

354/20

355/21

356/22

357/23

358/24

359/25

360/26

361/27

362/28

363/29

364/30

365/31

Date to right of slash (/) is day of month.

Initial Inputs

The following description of parameter inputs is given in the same order
as that of the input cards shown in table 11-15.

209

Table 11-15. — Parameter file for erosion/sediment yield component

Initial General Parameter Inputs

Card 1-3. TITLE ()

TITLE Three lines of 80 Characters each for alphanumeric
information to be printed at the beginning of the out-
put, format (20A4)

Card 4. BDATE, FLGOUT, FLGPAS, FLGPRT, FLGSEQ

BDATE The beginning date for simulation. It must be less
than the first storm date (SDATE) . (Julian date), e.g.
73000

Card 5.

FLGOUT 0 for annual summary output

1 for monthly and annual summary output

2 for storm by storm and both types of summary output

3 for a single storm and detailed output by segments

FLGPAS 0 if no file should be created for the Chemicals pro-
gram

1 if the program should create a file for use by the
Chemicals program

FLGPRT 0 for the particle specifications to be computed with
default values
1 for the particle specifications to be read in

FLGSEQ Execution sequence of erosion submodels:

1 - overland

2 - overland-pond

3 - overland-channel

4 - overland-channel-channel

5 - overland-channel-pond

6 - overland-channel-channel-pond

FLGSEQ is used to decide whether certain groups of cards
should be read in. Cards 9-11 are always read, and only once.
Cards 12-15 are only read when FLGSEQ is greater than or equal to
3, and they are repeated for a second channel if FLGSEQ is 4 or
6. Cards 16 and 17 are read if FLGSEQ is 2,5, or 6, and they are
never read more than once.

KINVIS, NBAROV, WTDSOI , KR, NBARCH, YALCON

If a default value is to be used, leave that position on the card
blank. Otherwise enter the desired value. If all defaults are
assumed, insert a blank card.

210

Table 11-15. — Parameter file for erosion/sediment yield component—continued

KINVIS
NBAROV

WTDSOI
KR

NBARCH

YALCON

Kinematic viscosity (ft /sec), e.g. default 1.21E-05

Manning's n for overland flow over bare soil, e.g.
default 0.01

Weight density of soil (lbs/ft3), e.g. default 96.0

Soil erodibility for erosiop-by concentrated flow
((lbs/ft^ sec) (1/lbs/ftV ) e.g. default 0.135

Manning's n for channel flow over bare soil e.g.
default 0.03

Yalin constant for sediment transport, e.g.
0.635

default

Card 6. SOLCLY, SOLSLT, SOLSND, SOLORG, SSCLY, SSSLT, SSSND, SSORG
SOLCLY

SOLSLT

SOLSND

SOLORG

SSCLY

SSSLT

SSSND

SSORG

Fraction of clay in the original surface soil layer
exposed to erosion, e.g. 0.14

Fraction of silt in the original surface soil layer
exposed to erosion, e.g. 0.20

Fraction of sand in the original surface soil layer
exposed to erosion, e.g. 0.66

Fraction of organic matter in the original surface soil
layer exposed to erosion, e.g. 0.01

2
Specific surface area of clay particles (meters /gram

of soil) , e.g. 20.0

2
Specific surface area of silt particles (meters /gram

of soil) , e.g. 4.0

2
Specific surface area of sand particles (meters /gram

of soil) , e.g. 0.05

Specific surface area of organic matter particles
(meters /gram of organic carbon), e.g. 1000.0
(organic carbon = organic matter/1.73)

The fractions of clay, silt, and sand should total 1.0, with
the organic matter being a fraction, of the total of organic
matter and soil particles.

If the specific surface area values are left blank the model
defaults to 20.0, 4.0, 0.05, and 1000.0 for clay, silt, sand, and
organic matter respectively.

211

Table 11-15. — Parameter file for erosion/sediment yield component—continued

If the particle specifications flag (FLGPRT, card 4) is 0 then no card 7
or card 8's will be read, and the number of particle types (NPART, card 7)
will be calculated.

Card 7. NPART

NPART The number of particle types, e.g. 5
Card 8. DIAM, SPG, FRAC, FRCLY, FRSLT, FRSND, FRORG

[Repeat card 8 for each particle (NPART, card 7)]

DIAM

SPG

FRAC

FRCLY
FRSLT
FRSND
FRORG

Particle diameter (mm) , e.g. 0.030
Specific gravity of particle (g/cnf

e.g. 1.8

Fraction of sediment detached that is made up of this
particular particle type, e.g. 0.50

Fraction of particle made up of clay, e.g. 0.3

Fraction of particle made up of silt, e.g. 0.5

Fraction of particle made up of sand, e.g. 0.2

Fraction of particle made up of organic matter
0.02

e.g.

The sum of the fractions for clay, silt, and sand should
equal 1.0, with the organic matter being a fraction of the total
organic matter and soil particles.

Initial Overland Flow Inputs

Card 9. DATOV, SLNGTH, AVGSLP, SB, SM, SE, XIN(3), YIN(3), XIN(4), YIN(4)

DATOV Area represented by overland flow profile (acres) , e.g.
3.2

SLNGTH Slope length of representative overland flow profile
(ft) , e.g. 206.0

AVGSLP Average slope of representative overland flow profile
(ft/ft) , e.g. 0.027

SB Slope at the upper end of profile, e.g. 0.020

SM Slope of mid-section, e.g. 0.0380

SE Slope at the lower end of profile, e.g. 0.024

212

Table 11-15. — Parameter file for erosion/sediment yield component— continued

XIN(3) Distance from top of slope where mid-uniform section
begins (ft) , e.g. 98.0

YIN (3) Elevation above lowest point where mid-uniform section
begins (ft) , e.g. 3.5

XIN(4) Distance from top of slope where mid-uniform section
ends (ft), 156.0

YIN (4) Elevation above lowest point where mid-uniform section
ends (ft) , e.g. 0.0

When simulating a uniform slope SB = SM = SE = AVGSLP;
XIN(3) = XIN(4) = SLNGTH; YIN(3) = YIN(4) = 0.0

Card 10. NK

NK Number of slope segments differentiated by changes in
soil erodibility factor, e.g. 1

Card 11. XKIN(I), KIN (I), ... for 1=1 to NK (card 10)

XKIN(I) Relative horizontal distance from the top of the slope
to the bottom of segment I, e.g. 1.0

KIN (I) Soil erodibility factor for slope segment just above
XKIN(I) (tons/acre/English EI) e.g. 0.23

The order of the following cards depends on the execution sequence
(FLGSEQ, card 4). In some cases the following cards (12-17) won't be used,
e.g. FLGSEQ = 1, or there may be two sets of channel inputs (12-15) and a pond
(16,17) , e.g. FLGSEQ = 6.

Initial Channel Inputs

Card 12. NS, FLAGC, FLAGS, CONTL, SECTN

NS Number of channel segments differentiated by changes in
slope, e.g. 5

FLAGC Flag that indicates channel shape:

1 - Triangular channel

2 - Rectangular channel

3 - Naturally eroded channel

FLAGS 1 for program to use curves for slopes of energy grade-
line, (friction slope)

2 for program to assume friction slope equals channel
slope.

213

Table 11-15. — Parameter file for erosion/sediment yield component—continued

Card 13.

Card 14.

Card 15,

CONTL 1 if critical depth controls depth in outlet channel

2 if uniform flow controls in the outlet channel

3 if the program should use the maximum of 1 and 2

4 if the program should use a rating curve for control
depth at outlet.

D = RA (Y - YBASE)
Q(ft /sec), Y and YBASE (ft),

SECTN 1 if the shape of the outlet channel is triangular
2 if the shape is rectangular

SIDSLP, BOTWID, OUTMAN, OUTSLP, RA, RN, YBASE

SIDSLP Side slope of a cross-section of the outlet control
channel, expressed as horizontal to vertical, e.g. 20.0

BOTWID Bottom width of the outlet control channel (ft), e.g.
10.0

OUTMAN Manning's N for the outlet control channel, e.g. 0.030

OUTSLP Slope of the outlet control channel, e.g. 0.002

RA Coefficient in the rating curve equation e.g. 2.41

RN Exponent in the rating curve equation, e.g. 2.25

YBASE Minimum depth for flow to begin (ft), e.g. 0.0

LNGTH, DATCH, DAUCH, Z

LNGTH Channel length (ft), e.g. 371.0

DATCH Total drainage area of channel at lower end of channel
(acres) , e.g. 3.2

DAUCH Drainage area above upper end of channel (acres) , e.g.
0.2

Z Sideslope of channel cross-section, expressed as hor-
izontal to vertical, e.g. 20.0

If the channel shape flag (FLAGC, card 12) is a 2 or 3,
enter the value for Z that most closely approximates the channel
shape.

TX(I), TS(I), . . . for 1=1 to NS (card 12)

TX(I) Distance from lower end of the channel to the bottom of
segment I (ft) , e.g. 0.0

TS(I) Slope of segment directly above TX(I), e.g. 0.024

214

Table 11-15. — Parameter file for erosion/sediment yield component—continued

Initial Pond Inputs

Card 16. CTL, PAC

CTL 1 for pipe outlet control as typical of impoundment
type terraces
3 when the orifice coefficient (C, card 17) is read in

PAC 1 for program to calculate coefficients for pond sur-
face area-depth relationship from user supplied parame-
ters for impoundment basin slopes.

2 for user supplied coefficients SA = FC(Y ), Where SA
= Surface area (FT ) , Y = depth (ft)

Card 17. DATPO, INTAKE, FRONT, DRAW, SIDE, FS, B, DIAO, C

DATPO Total drainage area above the pond (acres), e.g. 3.2

INTAKE Soil water intake rate within the pond, in/hr, e.g. 0.2

FRONT Embankment front slope, e.g. 0.2

DRAW Slope along channel draining into pond, e.g. 0.024

SIDE Slope of land at pond toward draw, e.g. 0.01

FS depth relationship, e.g. 9500.0

B depth relationship, e.g. 1.73

DIAO diameter of pipe orfice (in), e.g. 3.0

C Orifice coeficient, e.g. 3000.0

Updateable General Parameter Inputs

The remaining inputs to the Erosion program are updateable. The program
checks the dates (SDATE, card 1) from the hydrology file against the parame-
ters control date (CDATE, card 18). If the control date is less than the date
of the storm, the program reads in a new set of the updateable parameters. If
the program reads a blank in place of the control date (CDATE, card 18) the
program stops executing. The execution sequence flag (FLGSEQ, card 4) is used
to determine whether or not cards in this section are read as in the Initial
Inputs section. There are no updateable Pond parameters. The Overland flow
parameters are on cards 19-22, and the Channel parameters are on cards 23-29.

Card 18. PDATE, CDATE

PDATE First date that the following erosion parameters are
valid (Julian), e.g. 73138

215

Table 11-15. — Parameter file for erosion/sediment yield component—continued

The program doesn't read in the value for PDATE. PDATE
is only used as an aid in putting together the data
file.

CDATE Last date that the following erosion parameters are
valid (Julian), e.g. 73105

NOTE: A card 18. should always be the first card in a set of
updateable parameters.

Updateable Overland Flow Inputs

Card 19. NC, NP, NM

NC Number of slope segments differentiated by changes in
cropping management factor, e.g. 1

NP Number of slope segments differentiated by changes in
contouring factor, e.g. 1

NM Number of slope segments differentiated by changes in
Manning's N, e.g. 1

On the initial pass through the program, each of the "N"'s
should be at least 1 in order to read initial values for the
parameters. In subsequent passes, a blank "N" indicates no
change in the corresponding parameter from the previous update.
To skip reading a parameter, for example Manning's n, read in a
blank NM. If no new overland flow parameters are to be read,
card 19 should be left blank. Input cards for a parameter should
not be included in the data file when it's "N" is left blank.

Card 20. XCIN(I), CIN(I), ... for 1=1 to NC (card 19)

XCIN(I) Relative horizontal distance from top of slope to the
bottom of segment I , e.g. 1.0

CIN(I) Cropping management factor for slope segment just above
XCIN(I) , e.g. 0.26

Card 21. XPIN(I), PIN(I), ... for 1=1 to NP (card 19)

XPIN(I) Relative horizontal distance from top of slope to the
bottom of segment I , e.g. 1.0

PIN(I) Contouring factor for slope segment just above XPIN(I),
e.g. 1.0

Card 22. XMIN(I), MIN(I), ... for 1=1 to NM (card 19)

XMIN(I) Relative horizontal distance from top of slope to the
bottom of segment I , e.g. 1.0

216

Table 11-15. — Parameter file for erosion/sediment yield component—continued

MIN(I) Manning's N value for slope segment just above XMIN(I),

e.g. 0.03

Updateable Channel Inputs

Card 23. NN, NCR, NCV, NDN, NDS, NW

NN Number of channel segments differentiated by changes in
Manning's N, e.g. 1

NCR Number of channel segments differentiated by changes in
critical shear stress, e.g. 1

NCV Number of channel segments differentiated by changes in
shear stress for cover, e.g. 1

NDN Number of channel segments differentiated by changes in
depth from channel middle to the nonerodible layer,
e.g. 1

NDS Number of channel segments differentiated by changes in
depth from the channel side to the nonerodible layer,
e.g. 1

NW Number of channel segments differentiated by changes in
width, e.g. 1

On the initial pass through the program, each of the "N'"s
should be at least 1 in order to read initial values for the
parameters. In subsequent passes, a blank "N" indicates no
change in the corresponding parameter from the previous update.
To skip reading a parameter, for example channel width, read in a
blank NW. If no new channel parameters are to be read, card 23
should be left blank. Input cards for a parameter should not be
included in the data file when it's "N" is left blank.

Card 24. XN(I), TN(I), ... for 1=1 to NN (card 23)

XN(I) Distance from the lower end of the channel to the bot-
tom of segment I (ft) . e.g. 0.0

TN(I) Manning's n of channel directly above XN(I), e.g. 0.065

Card 25. XCR(I), TCR(I), ... for 1=1 to NCR (card 23)

XCR(I) Distance from the lower end of the channel to the bot-
tom of segment I (ft) . e.g. 0.0

TCR(I) Critical shear stress of channel directly above XCR(I),
(lbs/ft ), e.g. 0.40

217

Table 11-15.— Parameter file for erosion/sediment yield component—continued

Card 26. XCV(I) , TCV(I) , ... for 1=1 to NCV (card 23)

XCV(I) Distance from the lower end of the channel to the bot-
tom of segment I (ft), e.g. 0.0

TCV(I) Shear stress for cover stability for channel directly
above XCV(I) , (lbs/ft ) , e.g. 100.0

Card 27. XDN(I), TDN(I), ... for 1=1 to NDN (card 23)

XDN(I) Distance from the lower end of the channel to the bot-
tom of segment I (ft) , e.g. 0.0

TDN(I) Depth to the nonerodible layer in the middle of channel
directly above XDN(I) (ft), e.g. 0.33

Card 28. XDS(I), TDS(I), ... for 1=1 to NDS (card 23)

XDS(I) Distance from the lower end of the channel to the bot-
tom of segment I (ft), e.g. 0.0

TDS(I) Depth to the nonerodible layer along the side of chan-
nel directly above XDS(I), e.g. 0.33

Card 29. XW(I) , 1W(I) , ... for 1=1 to NW (card 23)

XW(I) Distance from the lower end of the channel to the bot-
tom of segment I (ft) , e.g. 0.0

1W(I) Width of channel directly above XW(I) (ft), e.g. 10.0

If the channel shape flag (FLAGC, card 12) is a 1 or 3, enter the
value for 1W that most closely approximates the channel shape.

Cards 19 and 23 must be included, depending on the execution sequence
(FLGSEQ, card 4), every time the updateable parameters are repeated. Cards
20-22 and 24-29 are included only if indicated on cards 19 and 23.

218

Table 11-15. — Parameter file for erosion/sediment yield component—continued

A sample partial data file for the Control Parameters follows. It will
help demonstrate the file structure.

EROSION PARAMETER DATA

EROSION PARAMETERS -

GEORGIA PIEDMONT

MANAGEMENT PRACTICE ONE

CON; ■ l-irn i' j nii'ii i :,;nj- in i!:;!ni i 1 1 i r,i i

73000

0.000

.000

0.000

0.140

0.200

.GGO

0.010

20.000

4.000 0.0501000,

.000

3.200

20G.000

.027

0.020

0.038

0.024 98.000 3,

.500

15G.000

0.000

1.000

0.230

20.000

10.000

.030

0.002

2.410

2.250 C.000

371.000

3.200

.200

20.000

0.000

0.024

.000

0.018

154.000

0.014 2G9.000 0,

.032

325.000

0.021

73105

1.000

0.2G0

1.000

0.030

0.000

0.0G5

0.000

0.400

0.000

100.000

0.000

0.330

0.000

0.330

0.000

10.000

73121

1.000

0.400

1.000

0.030

0.000

0.040

0.000

0.150

0.000

0.330

0.000

0.330

219

General Parameter Values

Starting Date— Set this value to zero if the model is used for a single storm.
For multiple storms, the date should be less than that of the first storm (1
day less is sufficient). The date is given by first specifying the last two
digits of the year (for example, 78) followed by the Julian date (for example,
094 for April 4, or 78094).

FLGOUT--This input determines whether the model runs for a single storm or for
a series of storms. A 0 selects multiple storms, but the output is limited to
annual summaries. A 1 selects multiple storms and gives output as monthly and
annual summaries. A 2 gives output for each storm as well as the summaries. A
3 is used when the model is run for a single storm. It gives additional output
of soil loss for each segment on the overland flow and channel elements. This
indicates areas in the watershed where intense erosion or deposition occurs.

FLGPRT--If the particle distribution is computed, it is computed from the pri-
mary particle size distribution of the original soil mass. Management and
other factors affecting aggregate sizes are not considered.

FLGPAS--Set to 0 if erosion/sediment yield estimates are not needed in other
computations outside of the erosion/sediment yield component. Set to 1, the
model writes date (Julian), volume of rainfall (in), volume of runoff (in), en-
richment ratio (specific surface area of sediment and organic matter to that of
original soil mass), sediment yield per unit area (tons/acre), and values input
into the program for DP, PERCOL, AVGTMP, AVGPEV, POTPEV, ACCSEV, and POTSEV.
These data are written into file 7, named PASS.

Sequence—The watershed is represented by a combination of such elements as
overland flow, channel, and pond, and the calling sequence of these elements.
Table 11-16 gives the permissible sequences.

Table 11-16. — Elements and their sequence numbers to represent main watershed

features

Sequence number Sequence of elements

!_________________ Overland.

2----------------- Overland-pond.

3_________________ Overland-channel .

4_________________ Overland-channel-channel.

5_________________ Overland-channel-pond.

6_________________ Overland-channel-channel-pond.

Before selecting the element sequence number, identify major features in
the watershed that affect erosion and sediment yield. An aerial photograph and
a site visit are especially useful. USGS topographic maps are generally too
coarse for this application. A representative overland flow profile, chan-
nel (s), and impoundment are used to characterize the watershed elements. This
characterization is discussed in later sections.

All watersheds are assumed to be composed of an overland flow element.
Natural topography causes overland flow to converge into major flow concentra-
tions on many farm fields. These few concentrations are readily distinguish-

220

able from the many rills that may exist on a field. The definition of a rill
becoming a gully when a rill can no longer be obliterated by tillage is not
workable, nor is the definition workable that a rill becomes a gully when it
exceeds a certain size. The critical factor is how rills behave hydraulically.
Removing a single rill has a negligible effect on the hydrologic-hydraul ic
response of the watershed, whereas removal of a single flow concentration has a
major effect. Flow concentrations are easily identifiable with a site visit to
a field tilled immediately before a major storm. In fact, site visits to typi-
cal fields before using the model are very helpful.

Other flow concentrations may exist besides these natural ones. Examples

include terrace channels, grass waterways, and diversion ditches. A ridge

develops around many fields, which often collects overland flow and causes a
flow concentration along the edge of the field.

Impoundment terraces obviously pond water and are represented by a pond
element. Other types of ponds are formed by natural depressions, roadways with
pipe culverts, and other structures. A ridge and dense grass around the edge
of many fields may pond runoff, causin-g considerable deposition near the edge
of the field.

Obtain a map of the area to be modeled and identify the watershed bound-
ary. Next, identify the channel and pond elements within the watershed. Only
a single overland flow element may be called, which always is called first, and
only a single pond element may be called, which always is called last. Flow
through a series of ponds cannot be modeled. Typical examples are:

1. If an estimate of erosion on the overland flow areas alone is needed
or if the field area is a simple overland flow area adjacent to a
stream, only the overland flow element is called (CSEQ = 1) (Fig.
II-17a).

2. The study area may be a simple watershed with a single concentration
of flow down the middle, an overland flow section draining down a
ridge formed by a field boundary that directs the flow along the field
edge as concentrated flow, or an overland flow section cut off by a
diversion ditch (CSEQ = 2) (fig. II-17b).

3. The study area may be a watershed with a major main flow concentration
with several lateral flow concentrations feeding it, or it may be a
series of terrace channels feeding an outlet channel (CSEQ = 4) (fig.
II-17c).

4. If backwater is at the outlet for any of these situations, a second
channel with backwater outlet control is added to the sequence
(example 1 would become CSEQ = 3, or example 2 would be CSEQ = 4 (fig.
II-17e)).

5. For impoundment terraces, two options are available for delivering
flow to the impoundment. Overland flow goes directly to the impound-
ment (CSEQ = 2) (fig. II-17d), or overland flow first goes to a chan-
nel and then goes to the pond. Two channels may be involved, one in
the draw, and one along the terrace. Select parameters based on the

221

OVERLAND FLOW

(X$,0)

(I) OVERLAND FLOW
SEQUENCE AND SLOPE REPRESENTATION

OVERLANO FLOW
SLOPE REPRESENTATION

^-"•(0>V

JtX,,Y,l

AVERAGE

SLOPE -^ / 1 M,° SLOPE

/ y/(X4,Y4) _

OVERLAND

IMPOUNDMENT
TERRACE

CONCENTRATED FLOW

(2) OVERLAND FLOW
POND SEQUENCE

(3) OVERLAND FLOW
CHANNEL SEQUENCE

OVERLAND
s~ FLOW

OVERLAND FLOW

m 1

1 \

V \ TERRACE
r FLOW

JIM

CHANNEL FLOW . /

V. OUTLET
CHANNEL FLOW

: •■" /

POND AT —

FIELD OUTLET

(4) OVERLAND FLOW
CHANNEL-CHANNEL SEQUENCE

(5) OVERLAND FLOW
CHANNEL-POND SEQUENCE

Figure 11-17. — Schematic representation of typical field systems in the
field-scale erosion/sediment yield model.

one delivering the most flow. The model does not permit a combination
of both overland flow and channel delivery to the pond. The model can
be run assuming different delivery systems and averaging results.

6. The pond outlet is assumed to be outside the study area, which prohi-
bits analyzing ponds in a series.

7. If the sequence changes during the simulation period, the simulation
must be run in parts, stopping when the sequence changes and then

222

restarting.

Kinematic Viscosity--The model defaults to a kinematic viscosity of 1.21
x 10~5 ft^/s, the value for a temperature of 60° F. The value is assumed to
be constant for the duration of the simulation period. The value was chosen
assuming that most highly erosive storms occur in April and May. The value
should be selected according to the temperature when most erosive storms occur
if the model is being run for multiple storms. If it is being run for a single
storm, a value appropriate for the temperature at the time of the storm should
be selected. Table 11-17 gives kinematic viscosity values for a range of tem-
peratures.

Table 11-17 Kinematic viscosity for water over a typical range of tempera-
tures

Temperature Kinematic viscosity

(!£) (ft2/s x 105)

40 1.67

50 1.41

60 1.21

70 1.05

80 0.90

90 0.82

100 0.74

Manning's n for overland flow over bare soil --The default value is 0.01, which
represents smooth areas where broad overland flow has deposited sediment ( 5J .
Although the overland flow surface may be much rougher, a value for a smooth
surface must be input because it is used to compute the portion of the flow's
total shear stress that acts on the soil to cause detachment and transport.

Weight density of soil mass--This input is for the weight density of the soil
mass in areas of flow concentrations. Although tillage, soil, management, time
of year, and other factors significantly affect the density of the soil mass, a
constant density is assumed over the simulation period. The default is
96 lb/ft3 (1.54 g/cm3 bulk density), which is larger than typical for surface
soils. Flow concentrations typically occur in low areas and if erosion has
occurred, much of the original surface soil is gone, leaving a more dense
subsurface soil exposed. Compaction from farm equipment also is assumed to be
greater in these areas. Therefore, bulk densities typical of tilled surface
soils, especially after tillage, may be too small. The bulk density of the B
horizon is probably a good value. On 36 of the Indiana soils used in the soil
erodibility study by Wischmeier and others (13), bulk density of the B horizon
ranged from 0.97 to 1.70 g/cm3 with a mean of 1.37 g/cm3 and a standard
deviation of 0.15 g/cm3. The mean weight density was 85 lb/ft3. Table
11-18 is recommended for selecting a value.

Soil Erodibility Factor for Erosion by Concentrated Flow—Some soils are much
less susceptible to erosion by flow. Little information is available in the
literature that may be used to estimate soil erodibility due to flow. A value

223

of 0.135 (lb/ft2/s)(ft2/lb)1-05 was obtained experimentally in a rill
erosion study on a tilled silt loam soil (vol. Ill, ch. 11). For most applica-
tions, the default value is recommended. If the factor is varied, estimate the
first approximation of K from the soil erodibility nomograph of Wischmeier and
others (_13) and multiply by 0.39. This assumes Krcn = 0.135 (lb/ft2/s) for a
first approximation of K = 0.35 (ton/ac/EI). Use the default value for sandy
soils. Soil structure and permeability in the nomograph of Wischmeier and
others (13) are considered nonapplicable to erosion by concentrated flow.

Table 11-18 — Bulk and weight densities in areas of concentrated flow

Condition Bulk density Weight density

(g/cm3) (lb/ft3)

Loose. 1.20 75

Not subject to compaction 1.37 85

and tilled regularly with
primary tillage equipment.

Subject to compaction and 1.54 96

tilled regularly with pri-
mary tillage equipment.

Not subject to compaction 1.54 98

and not tilled regularly
with primary tillage equip-
ment.

Subject to compaction and 1.65 103

not tilled regularly.

Manning's n for Channel Flow over Bare Soil --The default value is 0.03, which
seems typical for nonvegetated earth channels, such as those channels where
flow concentrates in farm fields {I, 8). This value is also consistent with
Manning's n estimated from rill erosion experiments (vol. Ill, ch. 11). This n
represents the roughness for flow over a seedbed or a relatively smooth soil
that has eroded down to a nonerodible layer. Although the channel being analy-
zed is rough, covered with crop residue, or vegetated, a value for bare soil
must be input because it is used to compute the portion of the flow's total
shear stress that acts on the soil to cause detachment and transport.

Yalin's Constant--Ya1in's sediment transport equation contains a constant equal
to 0.635, which Yalin (14) obtained by fitting his equation to approximate
sediment transport data from natural stream channels. When the equation was
tested against overland flow data, the constant had to be increased to 0.88 to
give good results for sand and it had to be decreased to 0.47 for coal parti-
cles having a 1.6 specific gravity (2). This equation is for bedload transport
and may underpredict when a significant quantity of sediment is transported as
suspended load. The constant can be increased to account for the suspended
load transport capacity. No values are suggested, however. Use 0.635 unless
other validated values are available (vol. Ill, ch. 10).

224

Particle Description

Particle Distribution of Residual Soil

The action of clay, silt, sand, and organic matter are for the original
soil mass in the upper layer exposed to erosion and are based on the standard
USDA classification. Use soil survey information or soil tests to estimate
these values. The fractions are expressed so that the soil mineral fractions
total 1.00. The fraction of organic matter is expressed as part of the total.
These data are used to compute the distribution of particles of detached sedi-
ment if that option (FLGPRT = 0) is selected. With FLGPRT = 0 or 1 , these data
are used to compute enrichment ratios for the eroded sediment. (See vol. I, ch.
3 for procedures.)

Description of Sediment Particles

Two options are available in describing detached sediment particles and
organic matter. The first option is to use the assumed relationships described
in volume I, chapter 3. The second option is to input detailed information on
the particles. Required information includes particle distribution (diameter,
specific gravity, and fraction) of the sediment as it is detached, composition
of each particle type, the specific surface area of clay, silt, sand, and or-
ganic matter; and relation of organic matter to clay in the eroded sediment.

Only limited information is given for choosing input values for the second
option. A user that chooses the second option must research the required in-
formation on his own.

Sediment mixtures composed of up to twenty particle types are allowed with
the second option. A type is a specific combination of size and density.
Although clay-sized particles can be specified, flocculation or dispersion is
not considered. If water or sediment chemistry causes flocculation or disper-
sion, identify this potential, enter appropriate particle size and density, and
interpret the model results accordingly.

Many soils erode as aggregates, which are conglomerates composed of
primary sand, silt, and clay particles having specific gravities less than the
density of primary particles. Particle sizes are functions of soil properties,
management, cover, and detachment by raindrop impact vs. detachment by runoff.
Based on Young's analysis!/ of available data, the following particle types are
suggested in table 11-19.

Although the classes in table 11-19 are rather broad, do not deviate
greatly from them except where the soil is poorly aggregated and aggregate
stability is low. Since most agricultural soils erode as aggregates,
especially in the midwest, size distribution of primary particles should be
used directly as the eroded particle distribution only if the soil is totally
nonaggregated when it erodes.

2/Personal communication with R. A. Young, USDA-SEA-AR Morris, Minn.

225

Table 11-19. — Soil particle types

Fraction in

Condition

Size

Density

detached
sediment

Particle

(mm)

(g/cm3)

Soils with ratio of

0.002

2.60

0.05

Primary clay.

silt to sand and clay

.010

2.65

.08

Primary silt.

> 0.5 (sa 15%, si

.020

1.80

.50

Small aggregate.

60%, cl 5%).

.500

1.60 .31

Large aggregate.

.200

2.65

.06

Primary sand.

High clay soils (sa

.002

2.60

.10

Primary clay.

10%, si 40%,

.010

2.65

.06

Primary silt.

cl 50%).

.075

1.80

.57

Small aggregate.

1.000

1.60

.25

Large aggregate.

.200

2.65

.02

Primary sand.

High sand soils (sa

.002

2.60

.02

Primary clay.

75%, si 15%, cl

.010

2.65

.02

Primary silt.

10%).

.030

1.80

.16

Small aggregate.

.200

1.60

.20

Large aggregate.

.200

2.65

.60

Primary sand.

Factors that reduce rill erosion in relation to interrill erosion seem to
increase the amount of primary particles and small aggregates in the fine size
range. Particles from rough surfaces, vegetated surfaces, and flatter slopes
tend to be smaller.

Data for these effects are limited. The figures and the discussion on
their use are given only to indicate the effects, which will help interpret
results. To adjust values in table 11-19 for slope, read off a factor value
from figure 11-18. Multiply the fractions for the large particles, (that is,
large aggregate and primary sand) by the factor and increase the other fraction
in equal proportions to account for the reduction so that the total of the
fractions is one.

Use figure 11-19 to adjust for cover and the effect of consolidation.
When soil lays exposed to consolidating traffic, it becomes more resistant to
rill erosion and particle size is believed to decrease. The consolidation
curve represents the change over the growing season or the effect of notillage,
which is assumed not to change over time. Primary tillage lies in between.
These curves do not account for roughness from primary tillage and its effect
on transport capacity, which must be considered in the hydraulic roughness in
put. Figures 11-18 and 11-19 are extrapolations from Young's review2/of avail-
available data on particle size, and should be used carefully. In most prob-
lems, the effects described in these figures can be neglected.

-f Op Cit

226

5 10 15
SLOPE (PERCENT)

;'<o

Figure 11-18. — Adjustment factor for mul-
tiplying fractions of large parti-
cles (large aggregates and sand) to
account for effect of slope on par-
ticle size. (Note: Curves are
speculative.)

Only limited information is given on selection of information for particle
characteristics to help the user obtain detailed information. Information on
specific surface area is available in texts on soil physics. Note that speci-
fic surface area is used for organic carbon, rather than organic matter. If
values are not input for specific surface areas, the model devaults to 20, 4,
0.05, and 1,000 mP/g, respectively, for clay, silt, sand, and organic carbon.

■X i

The specific surface area of 20 m^/g is typical of kaolionitic clay,
illonite clay may be as high as 800 m2/g.

Montmor-

Particle composition is used to compute specific area and enrichment ratio
of specific surface area. Aggregates are made up of organic matter, clay,
silt, and sand. The specific surface area of the aggregate depends on the com-
position of these basic components. Little information on aggregate composi-
tion is available. The equations given in volume I, chapter 3 may be used as
initial approximations.

Particle composition is not used to compute transport capacity and deposi-
tion of the particle types. Consequently, sediment yield estimates are accu-
rate regardless of the accuracy of the input of particle composition.

Overland Flow

The overland flow element represents typical overland flow conditions on
the watershed. After a representative land profile is selected, values must be
identified for the erosion variables and their relative locations along the
representative profile. The model uses averages in the direction perpendicular
to the downslope direction, that is, along the contour but not along the slope.
Thus, spatial effects along the profile can be considered with the model.
Variations in cropping practices over the field cannot be analyzed with the
model except for changes in cropping practice along the profile, such as strip
cropping and grass buffer strips. The model assumes uniformity along the con-
tour.

227

•CONVENTIONAL SEEDBED FOR
CORN AND SOYBEANS

• CHISEL, PLOW, DISK
PRIMARY TILLAGE

CONSOLIDATED
OR NO-TILL

40 60

COVER (PERCENT)

Figure 11-19 — Adjustment factor for multiplying
fraction of large particles (large aggregates
and sand) to account for effect of cover and
"consolidation" (soil disturbance) on parti-
cle size. (Note Curves are speculative.)

Location on the slope of such features as highly erodible soils dramatic-
ally can effect sediment yield. A sediment delivery ratio concept or a P fac-
tor concept from the USLE, except for contouring, is not used. Both sediment
delivery ratio and P factors are highly variable from storm to storm.

Nonupdateabl e Parameters

Overland Flow Area—This area of the watershed is represented by the overland
flow element. It usually equals the total watershed area.

A modification of the USLE is used to compute detachment on the overland
flow element. The modified equation uses soil erodibility, crop stage soil
loss ratio, and contouring factors from the USLE without change. The next step
is to assign values for these parameters and other detachment -transport param-
eters along the land profile.

Slope 1ength--0n simple rectangular areas, slope length is the distance from
the point that overland flow originates to where it reaches concentrated flow.
On typical midwestern fields, slope lengths seldom exceed 300 ft unless flow
is constrained by tillage marks, crop row ridges, graded furrows, or formed
channels. Do not use USGS contour maps to estimate slope length; they usually
give excessively long slope lengths.

On simple areas, such as between terraces, the slope length is the typical
distance between terraces. On more complex areas, the method of Williams and
Berndt (11) may be used. This contour-extreme point method requires a contour
map with the closest contour intervals available.

When a flow concentration crosses a contour, the contour comes to a point

228

generally in the direction of the watershed divide (fig. 11-20). These are
called extreme points because they are local maxima in an uphill direction.
Three contours, LC25, LC50, and LC75 are located at 25, 50, and 75% of the
total watershed relief, and their lengths are determined. Next, the length is
measured around the base of the LC contour (LB in fig. 11-20). Slope length is
given by:

X = (LC50 ' LB)/(2EP • (LC502 - LB502) 1/2) [II-7]

where EP = number of extreme points.

— ^ CONTOUR

LB-CONTOUR BASE

/ EXTREME

\ j / POINTS

Figure 11-20. — Sample watershed show-
ing contour extreme points and
base contour.

The resulting slope length should be inspected to determine if it is rea-
sonable. This method breaks down as the watershed approaches a simple plane
such as the area between terraces.

A subjective approach also may be used to determine such characteristics
of slope as length and steepness. The watershed is divided into 10 to 15 areas
approximately equal in area. Flow (stream) lines are drawn perpendicular to
the contour lines. Slope length for each stream line is the distance from
where overland flow originates to where it reaches concentrated flow, such as
that in waterways or drainageways in farm fields. The slope length used in the
model is the average of these lengths. Slope parameters to be discussed also
can be evaluated for each flow line and averaged.

229

Average slope steepness of representative overland flow profile—Determini ng
steepness is simple for simple land forms. The methods of Williams and
Berndt's (11) are recommended for complex watersheds:

S = 0.25 Z (LC25 + LC50 + LC75)/DATWS [II-8]

where Z = difference between an elevation of the highest point in the watershed
and the elevation of the outlet and DATWS = total area of the watershed. The
result should be inspected for consistency and reasonableness. The preceding
subjective approach may be used as an alternate.

Profile midsection--A typical overland flow profile is identified for the
watershed. The length and steepness of its slope have been determined. The
following steps fill in its shape, soil, and cover characteristics.

Field profiles occur in a variety of shapes, such as those shown in figure
11-21. The elements of slope profile used by the model to represent the actual
field profile are identified in the complex, convex-concave slope in figure
11-21. Read from left to right to identify slope elements and to name a slope.
A slope is complex if it has convex and concave elements. It is a complex,
convex-concave slope if the first curved element is convex. Values are
assigned one at a time to the variables used to describe slope shape until all
variables required to describe the slope have assigned values. Any remaining
variables are assigned the last value appropriate for that type of variable.

UNIFORM UPPER SECTION (SL0PE = SB)

CONVEX UPPER SECTION

UNIFORM MIDSECTION (SLOPE = SM)

AVERAGE SLOPE

CONCAVE LOWER SECTION

UNIFORM LOWER SECTION (SLOPE = SE)
0 SLOPE LENGTH

COMPLEX SLOPE '. CONVEX -CONCAVE

SLOPE LENGTH

COMPLEX SLOPE \
CONCAVE-CONVEX

'0 SLOPE LENGTH
SIMPLE CONCAVE

0 SLOPE LENGTH
SIMPLE CONVEX

"0 SLOPE LENGTH
SIMPLE UNIFORM

Figure 11-21

-Slope shapes that can be analyzed with the erosion/sediment
yield model .

230

In the coordinate system, x = 0 at the origin of overland flow where y =
maximum elevation and y = 0 at x = slope length.

If a midsection exists, it must be located and its coordinates must be
determined. If an upper convex section immediately changes to a concave
section, no miduniform section exists. In this situation, coordinates of the
lower end of the convex portion and the upper end of the concave portion are
set equal to each other and equal to the coordinates where the two curves meet.
The slope at this point must be specified later as SM. For a simple slope,
coordinates of the midsection will be those at the upper and lower ends of the
miduniform segment, if it exists. If the miduniform section does not exist,
the coordinates of both ends of the miduniform section are set equal to the
coordinates of the end of the slope, that is, x = slope length; y = 0.

Slope at upper end--The slope at the upper end, SB, is the slope of the upper
end of the uniform segment, if it exists. If the uniform segment does not
exist, SB is the slope at x = 0.

Slope of midsection--The slope of the midsection, SM, is the slope of the
miduniform segment, if it exists. If not, it equals slope of the land profile
where the upper and lower sections meet. On simple convex or concave slope, it
is slope of the land profile at x = slope length.

Slope at lower end--The slope at lower end, SE, is the slope of the lower end
of the uniform segment, if it exists. If not, SE is the slope of the land
profile at the end of the profile where x = slope length. For uniform slopes,
SB = SM = SE . For simple concave and convex slopes, SM = SE. On convex
(simple or complex) slopes, SM is greater than the average slope and less than
the average slope for concave (simple or complex) slopes. The upper slope, SB,
is less than the average slopes for convex slopes and greater than the average
slope for concave slopes. On complex, convex-concave slopes, SE is less than
the average slope, while on complex, concave-convex slopes, SE is greater than
the average slope. Failure to satisfy these conditions may cause fatal errors
during the program execution (for example, dividing by zero or raising negative
numbers to a power) .

Soil erodibility— Some soils are more erodible than others. The erodibility of
a soil is expressed by the soil erodibility factor. Values for this factor may
be estimated from a soil erodibility nomograph (fig. 11-22) (12). Using this
nomograph requires a mechanical analysis of the soil to determine sand (0.1-2.0
mm), very fine sand, silt, clay (USDA classification), and organic matter frac-
tions. Soil survey classification for soil structure and permeability also is
required. Local offices of the USDA-Soil Conservation Service usually can pro-
vide soil erodibility values for local soils.

Identify the relative position (distance from top of slope/slope length)
along the slope where the factor changes. The factor is assumed to be constant
over the slope segment just above the point of change. If the entire slope has
a single factor value, 1.0 is entered for relative distance.

To illustrate input, assume a slope length of 200 ft and K = 0.4
tons/acre/EI for the first 150 ft, and 0.2 ton/acre/EI for the last 50 ft. The

231

,\\

p —

- J

<
in

UJ
\0.

o o
* E • w

\cm\

\ \\

very *lo
slow
slow fo
modera
mod. to
rapid

very fine gronulor

fine granular
5 med. or coarse granular
» blocky, platy or massive

* SOIL STRUCTURE

\N^\

V\V

» K) CM —

\ \. \

\ ^v

N \

N \ s

»*•

\
— *■■

s\\

* 3

o o c

h- to U

D N0I1

VINIXO
5 C

c
dddV

3 C

isau
5 <

5 A c

- ! <

3 C
3

■> O C
r ro c

y Ainiaiao

5 O O
J3-TI0S

1
8s A

a O Q.

10 o^ *

■ ill

~ O c

— • •

O CM 3

Ji? *•

,-«

1 •!

a o"tj •

° ~ C 3

\ \

t /: ,

-4^

\ \

/ j/

• £,

\ V

/ ; j
/ / /

/ /'

' z
<
00

E
E
q

5/%

'/ A

r-
Z
UJ

/
/ /

O / //

'fa

/vX

UJ
CL

// v

' ' a "

a -
uj £

: 'cl

£2

." •

• o «:;

d'66'66'o'<

- cm to * u"> to r
QNVS 3NI3 AH3A + HIS 1N3

r> o o c

OD CT> <

-M +J -C

C C "3

' — -

CD CD CD

U cj E

s- s_ s-

■^

C> QJ OJ

CL Q. CL

' — '

C CD

S-

73 S «

a o

X <D

C
4->

o^-^

>4-

c: ro -o

O
O

(71 | "3

O r > "

LO g

S-

s- +

4->

.c ^^

+->

-2 ro

. '

x: csj

•i—

CL.— 1

c •

ra ^— >

cnool

cr,^-

s- --1

O^r

O —

E I

o o

+J -C

CI rH

C 4->

QJ -i-

u E

-M^r

s- oo

1 — •

CLX3

-q s:

II ro

•a ^-i

<t3 S-

o •

i_ C\J

^-^ QJ

i ii

o E

i —

jz:

•r- i^

-(-> u

c to

oo CD

OJ -r-

O 3

CJ .—i

S-

-C

CJ E

h- to

CL O

I 'r~

CD U-

• c:

CNJ O

CM -r-

•

1 -l->

^h ro

.«— « 00

i— « =3

oo ro

CT4- r—

<d a

> U

232

o>— o. O Ol ■» CO o-

2 s

O CM "t I*. "» ">

£2»8

o — cm <">

a =

■© o. cm >o o. cm _ n 10 n n « n

IN [ t- JE CM CI C»» C"> CM CM CM CM

-O O M> o«l»1 - O «1 N NOBS

£SS 22S3 S«33 :28S

23S 228S S33S = 2 8 S

I ! ! ! i ! s?«; : !

! ! ! i I I II I I II ! I II Mil

§§§§ §§§§§§§£§§§§ !§§§

I 5 UJ c

°KS'l

* a

a "8

-• a

3 _

S 5

a a a a

O — I o —

! I ?

S3S
PI S 5

I ! '

! I ?
9 5 3
33S

n a S

! ! I

x o "- i o •*■

TJ T> TJ "0

3^tO<o n a a o — n n t

nso

ol2.

Us]

a -5 „ ■*

5 I- I-
c > >

0 0 0

s I- I-

5 0 o

g = § c

Z "a-o^

ORN
o-ti//
3 to
1 to

c a. a

?ss

- 9 CM

= = o m

« >-

s SU

ssss

O O — CM O O — CM

S33J

as :

O K K JJgN
CO OK O O O —

§s

•o«o>-

O O cm -O r^ ao o o

= &

3^79
RC553

3S&3

CO O — <-> CO t^o — ovoo.

» CM -» N, —

I I N> «t «1 "1 CK.- . ! d P» | : ■ i CM ■» t». —

' n n n n « o r> <•> — — — cm

$s?s ssss as as ss?s «r,^«,cMZ.cM

3353 gsKR f5SS? 8 3 3S Nn,HD22SS
SSSfc 3223 SS5S 'QRfiS """"S^S

3533 2fcRfc I i i i i I I ; : i i

: I i , I ! ; 22" S82S8SS8 833388 833888

o „

at £

s &

TJ TJ
1"

-o r-. co o. o

J5SRS

SSSI^S S?;?g?

233

_aaa33S35 r r rr r £ l £ £ £ £

-o a a s s 2 2 a sssssarr -.-.«.«.««-okkkn« = k«o.o «o^n«»«.22-o«.k«. » 2

OJ — —

to Cn Pi H W . W i I W - ~^ — — — #- — ^ — — — — — — — — — — p— — ib CS •—

cushsss s;?s;?rr s 2 i 22a as a sss 9 s ssa ?S22«» 3a8S2:2 s 2

0 -' ~

YsaSSftSSS « * 3 fc 5 3 r _~ e:2sSSSa?!SS?S5!SS? BS{5S2r359sS£2 3 3

1 w ~

c?9SSiSR5 S33SRfcrr 2 2 a a a S ft 9 « 3 $ a R 3 8 3 8 S 9 3 S 8 2 S 3 8 9 a 3 £ 3 3

§ ■ ! 1 ! : sassaftrr 1 i 1 1 1 1 1 1 1 i 1 1 ! as*? a 1 1 1 1 I 1 1 1 1 1 1 3 3

2SS822fS8 Ml ! ! I 888388 382 882 ill 2 S 3 8 8 8 "■ 2 3 8 3 3 3 1

"5. So * 2 j

§■1 ? i. . is :- 1 - j]

O .c < . w Z. 3 -o • "> • a 2 o ! o -

e ^ =S -' "o 3 2 2 o 6 -2? S1- 2 5 "= * " £

v> ? - 52 2 * 5 t 8 3 5 " ° I .s u 2 * 4 £ ?

1i * » <v .- 5SM "8° <s 1 |1|8.

;r, .« a " *■ s - z -a "St z° "■ joo

£ - *? ? s J 5 * * * < s * jluS

oj22^22ssa ass sas asaaas ssa sa3 ^ 3333 3^338* B3338&8 a s

J_, ..-. „,.._._,-- __,».-_.__, — _ ^ _ _ ^„ _ „ _ ^ ^ ^ — , — „,.,.

^n«««c(n Snnnn nfill - — — -— — - NNrtri n « n p! cm <n <■»

I«»-'««oonin-n««in 9 99 >> 9 9 9 9 <aofsp>o.-esr. - cm ■»

^ - , „ . --Ml - <N - - - - - - - --- ".""*"

al°2*f * 22;«"< «»«< 9 99 ^2° °° -»< :nj< n«

"C " - - <S - - —

o 2 2 a s a s s s a 9 °° ^25 °° -° -° 22

— J-rt^->oiN«o<no.-o-iocM «? -. "T h ^. <►. 99 o-»o>omi>.<n<o cmo.

(/)---««« -nnSpi ««ni — — — — — — — c<— — c*<m ^^ —

.r»«o.t-)0>'Oi^— "Om— nrvw>« n. "". T ^f>°'. ^9 ««ns <o<- r>-o —

l/l— — — CNfJrt— ««nt MNfit — — — — --«n -non i c-<

"Z ■ ! ; : i i : : : : : i i : i i n ill! mm in

^SS^SSS 83SS2 3SS2 ii S8SSS838 9SS

Is ! I ! ' ' ! ' ! »SSSI9SS! SS!

*0 |a | ! "S I- " Is K S_

U|«*?A « 8 » - *I,°8 & 2

O -a°r.E-°£.|§.2.2.§8-2i8 " s

°^ ■ A3 I'* E-S = ; cQO ; cQ •? c ^ c

-* 5 * ii ji§o<s|5 !=..§:=§ I

- * -J *: J - i">**l-**~> > = *

i— o-a^-ffj-;^-; g

-Q Q B C M >

l_3333%« SH8 33 Sfc8& 3 53 333 33 HSkR^R s^r

234

2 t;
s «

"8 1 *

i 13 I

o> o o

I n

* = 2

^ 2 _s

•s ■ s

a .E ^

2 °

J ■;

"D -

,1J

t i. TJ >■

o> "* £

Jlj

8" 8

s-*_8

8 j

OS 0

"5
5 g

2 =
| f

i 9

* o> o

lis.

« .5 a •>

* ^ 8. ~
E - -S E

_2 o
> I

1 8-u

s "" i

en £K o

I* i

-S * 0 c

-S a

P ?

£ 2 ~

O _2- u

a o 5

-o o S

s §

c -

"a o

E
— =>

0 H

o o

o £

S 0

-£ ft

- O

■£ c

•* 8

| ft

E 0

I £

T> .1

3 U

-5 °>

1 s>

1 *5

S 1= i 5 i

y a, "
5 * «

' -2

«, ^ - _o

Z o

o _

9 £

.o ft

-D *

tl «C O -o

I* 4

E "->

U o- u «.

235

Table 11-21. — Approximate soil loss
ratios for cotton (12)

Expected final canopy percent cover:
Estimated initial percent cover from defc

65 80

30 45 6

Soil loss ratio

COTTON ANNUAL11
J. ..None.

Defoliatic

Feb

Mar

Cot Rd

Rd 4 20 percent cover vo
Rd & 30 percent cover vo

2 Chisel plow soon after cot ha

Chiseling to Dec. 31
Jan. 1 to sprg tillage

3... .Fall disk after chisel:
Disking to Dec 31
Jan. 1 to sprg tillage

4 Chisel plow Feb Mar, no prior

Cot Rd only

Rd 4 20 percent vol veg

Rd 4 30 perc-nt vol veg

5 Bed ("hip") feb-Mar, no prior

Cot Rd only

Rd 4 20 percent vol veg

Rd 4 30 percent vol veg

Split ridges & plant after hi

Disk & plant offer chisel ISt
Cot Rd only
Rd 4 20 percent vol veg
Rd 4 30 percent vol veg

Cropstage 1:
Cot Rd only
Rd 4 20 percent vol veg
Rd 4 30 percent vol veg

Cropstage 2

Cropstoge 3

6 Bed (hip) after I prior Ullage

Cot Rd only

Rd 4 20 percent veg

Rd 4 30 percent veg
Split ridges after hip (SB):

Cot Rd only

Rd 4 20 to 30 percent ve<
Cropstage 1:

Cot Rd only

Rd 4 20 to 30 percent ve(
Cropstage 2
Cropstage 3

7 Hip offer 2 pnor tallages.

Cot Rd only
Rd 4 20 30 percent veg
Split ridges after hip (SB)

8 Hip offer 3 or more tillages

Split ridges after hip (SB)

9 Conventional moldboard plow

Fallow period

Seedbed period

Cropstoge 1

Cropstage 2

Cropstoge 3

Cropstoge 4 (See practices

COTTON AFTER SOD CROP:

For the f.rst or second crop oft<
meadow has been turnplowed, multip
lines above by sod residual factors fr(

COTTON AFTER SOYBEANS:

Select values from above and multi

100

input card would be:
0.75 0.4 1.0

Updateable Parameters

0.2,

Selection of inputs thus far has
been discussed in the same order of the
inputs. Inputs to this point are fixed
for the simulation period. If they
must be updated, the run is stopped.
The input file is changed to the new
values, and a new run is started.

Initial constants for the channel
and pond elements are read before the
updateable inputs for the overland area
are read. The discussion continues
with the updateable overland flow vari-
ables.

PDATE, CDATE (Card 19) --This date, ex-
pressed as a Julian date, is the first
and last date that a set of parameters
is valid, including those of both over-
land flow and channel elements. Once
the storm date exceeds CDATE, the pro-
gram reads the next set of parameter
values. If zero is entered for the
number of data points for a parameter,
the program uses the most recent value
for that parameter. That is, only new
values are required for the parameters
that change.

56
47
42

46
38
19

1 Alternate procedure for estimating the soil loss ratios:

The rotios given above for cotton are based on estimates for re-

ductions in percent cover through normal winter loss and by the succes-

116

108
67

108
98
62

98
88
57

sive tillage operations. Research is underwoy in Mississippi to obtain

more accurate residue data in relation to tillage practices. This research
should provide more accurate soil loss rotios for cotton within a few

120
68

110

102

Where the reductions in percent cover by winter loss and tillage

operations ore small, the following procedure may be used to compute

soil loss ratios for the preplant and seedbed periods: Enter figure 6 with

the percentage of the field surface covered by residue mulch, move

vertically to the upper curve, and read the mulch factor on the icale

at the left. Multiply this factor by a factor selected from the following

tabulation to credit for effects of land-use residual, surface roughness
and porosity.

Productivitty No Rough Smoothed
level tillage surface surface

or gr.

1 the lost five

High 066 0.50 0.56
Medium 71 .54 .61

Poor .75 58 .65

of less than 1 percent should

See footnotes at right.

ilues for the bedded period on slopes _.

stimoted at twice the value computed above for rough surfac

»r vegetation.

Rd, crop residue, vol veg.

236

Table 11-22. — Soil loss ratios for con-
ditions not evaluated in table
11-20 (12)

Table 11-23. — Soil loss ratios (pet)
for cropstage 4 when stalks are
chopped and distributed without
soil tillage (12)

COTTON:

$•• table 5-A.
CROPSTAGE 4 FOR ROWCROPS:

Stalks broken and partially standing: Use col. 41.

Stalks standing after hand picking: Col. 41 times 1.15.

Stalks shredded without soil tillage: See table 5-C.

Fall chisel: Select values from lines 33-62, seedbed column.
CROPSTAGE 4 FOR SMALL GRAIN:

See table 5-C.
DOUBLE CROPPING:

Derive annual C value by selecting from table 5 the soil loss per-
centages for the successive cropstage periods of each crop.
ESTABLISHED MEADOW, FULL-YEAR PERCENTAGES:

and 2 and

Grass and legume mix, 3 to 5 t hay
Do. 2 to 3 t hay

Do. 1 t hay

Sericca, after second year
Red clover

Alfalfa, lespedeza, and second-year sericea
Sweetclover
MEADOW SEEDING WITHOUT NURSE CROP:

Determine appropriate lengths of cropstage periods SB,
apply values given for small grain seeding.
PEANUTS:

Comparison with soybeans is suggested.
PINEAPPLES:

Direct data not available. Tentative values derived analytically are
available from the SCS in Hawaii or the Western Technical Ser-
vice Center at Portland, Oreg. (Reference 5).
SORGHUM:

Select values given for corn, on the basis of expected crop residues
and canopy cover.
SUGARBEETS:

Direct data not available. Probably most nearly comparable to po-
tatoes, without the ridging credit.
SUGARCANE:

Tentative values available from sources given for pineapples.
SUMMER FALLOW IN LOW-RAINFALL AREAS, USE GRAIN OR ROW
CROP RESIDUES:
The approximate soil loss percentage after each successive tillage
operation may be obtained from the following tabulation by esti-
mating the percent surface cover after that tillage and selecting
the column for the appropriate amount of initial residue. The
given values credit benefits of the residue mulch, residues mixed
with soil by tillage, and the crop system residual.

Percent cover Initial residue (lbs/A)

Corn oi
Tilled

Sorghum

Soybeans

Mulch

Tilled

No-till in

Grain

rover1

seedbed2

No-till

seedbed2

corn rd3

Stubble'

—

1 Part of a field surface directly covered by pieces of residue mulch.

2 This column applies for all systems other than no-till.

3 Cover after bean harvest may include an appreciable number of
stalks carried over from the prior corn crop.

1 For grain with meadow seeding, include meadow growth in percent
cover and limit grain period 4 to 2 mo. Thereafter, classify as estab-
lished meadow.

Table II-24.— Factors to credit
residual effects of turned
sod1-7 (12)

by mulch

> 4,000

3,000

2,000

1,500

—

■14

—

'18

'25

Crop

Hay yield

Factor for

cropstage per

od:

SB and 1

Ton.

First yeor after

mead:

Row crop or

arc

3-5

0.25

0.40

0.45

0.50

0.60

2-3

.30

.45

.50

.55

.65

1-2

.50

.55

.60

Second year after

mead

. 3.5

.70

.80

.90

.95

.75

.85

.90

.93

1.0

1-2

.80

.90

1.0

Spring grain

2-3

—

.80

.85

.90

1.0

1-2

—

.85

.90

.95

1.0

Winter grain

2-3

.65

.75

.90

1.0

1-2

.70

.85

.95

1.0

1 For grain residue only.

WINTER COVER SEEDING IN ROW CROP STUBBLE OR RESIDUES:

Define cropstage periods based on the cover seeding date and apply
values from lines 129 to 145.

1 These factors an to be multiplied by the appropriate soil loss per-
centages selected from table 5. They are directly applicable for sod-
forming meadows of at least 1 full year duration, plowed not more
than 1 month before final seedbed preparation.

When sod is fall plowed for spring planting, the listed values for all
cropstage periods are increased by adding 0.02 for each additional
month by which the plowing precedes spring seedbed preparation. For
example, September plowing would precede May disking by 8 months
and 0.02(8—1), or 0.14, would be added to each value in the table. For
nonsod-forming meadows, like sweetclover or lespedeza, multiply the
factors by 1.2. When the computed value is greater than 1.0, us* as 1.0.

:37

Cover-management --Cover, tillage, stage of crop growth, and previous management
history greatly affect erosion. A grass-covered slope hardly erodes, while
erosion on a bare slope may be excessive. Similarly, a freshly prepared, fine-
ly tilled seedbed for corn is much more susceptible to erosion than it is imme-
diately following harvest.

I — 1 — 1 — 1 — 1—

■ ' '__■„-!-»--

„ "" ^-

— '

SMALL ,'
GRAIN^'

/ /

/ /corn

STALKS

1 L .,

0 1 2 3 4 5 6

MULCH (THOUSAND-POUNDS/ACRE)

Figure 11-23 — Relation of per-
centage of cover to dry
weight of uniformly distribu-
ted residue mulch. [From
Wischmeier and Smith (12).]

loss ratios in the tables of Wischmei
situations (but not all cases, such
continuous function in the model may
ratios at frequent intervals.

Values of the USLE crop-stage-soil -
loss ratio (SLR) appropriate for the his-
tory and present conditions on the field
are used to describe this factor. Tables
11-20 through 11-24 and Figure 11-23 for
soil loss ratios (SLR) were taken from
Wischmeier and Smith (12) , and should be
used to select cover-management factor
values. The C factor in the USLE is not
the same as the soil loss ratio. The C
factor is an average annual value that
integrates the variable soil loss ratio
and variable rainfall erosivity over the
year. This model does not use the aver-
age annual C factor.

The soil loss ratio describes ero-
sion characteristics at specific times
during the cropping cycles. The soil
er and Smith are step functions. In many
as at harvest), they are continuous. A
be approximated by reading new soil loss

Enter SLR values as a fraction rather than as a percentage, as shown in
the tables. Values are entered as (1) relative location where the soil loss
ratio changes and (2) the soil loss ratio just upslope of the point of change.
The following examples illustrate.

Example 1: Entire field is in continuous conventional corn at seedbed:

X*
1.0

SLR
0.8

Example 2: A 20-ft grass buffer strip is at the toe of a 200 ft slope

with corn at seedbed time. Total slope length is 220 ft. The relative

location for the first change is 200/220, or 0.91. The entered values might
be:

X* SLR x* SLR

0.91 0.80 1.00 0.03 (soil loss ratio for the

grass).

Example 3: The field is strip cropped with alternating strips of corn and
oats. Assume corn is 0 to 75 ft, oats is 75 to 125 ft, corn is 125 to 200 ft,

238

The entries could be:

SLR x SLR

SLR

0.10 0.8 0.80

1.0

0.10

and oats is 200 to 250 ft.
x SLR x

0.3 0.80 0.5

where x* = relative distance and SLR = soil loss ratio.

Contouring--Di recti on of tillage may significantly influence erosion and sedi-
ment yield. Contouring may store most runoff from small storms, greatly reduc-
ing sediment yield. For large storms that cause breakovers, however, sediment
yield can be greater than from uphill and downhill tillage. Following break-
overs, sediment yield from small storms increases. Unfortunately, contouring
factor values are defined poorly on a storm-by-storm basis. The values in
Table 11-25 are taken from the USLE and represent long-term averages. Future
refinements are needed greatly for this factor. Contouring loses its effec-
tiveness for long slopes. For slope lengths beyond those shown in table 11-25,
assign a contouring factor of 1.0.

Table 11-25. — Contour factor values and slope-length limits for contouring
[From Wischmeier and Smith (12)]

Contour factor Maximum

Land slope value length- _

(%) (ft)

1 to 2 0.60 400

3 to 5 .50 300

6 to 8 .50 200

9 to 12 .60 120

13 to 16 .70 80

17 to 20 .80 60

21 to 25 .90 50

— Limit may be increased by 25 % if residue cover after crop seedling
regularly will exceed 50 %.

Values in table 11-25 are for contouring typical of conventional farming
practices where row ridges are formed during cultivation after crops emerge.
Ridges are typically 4 to 6 in above the row middles.

Rows in many fields follow field boundaries rather than on the contour.
In one part of the watershed, rows may be directly uphill and downhill while in
another part they may be on the contour and somewhere in between for the rest
of the watershed. Effectiveness of contouring is partially due to deposition
in the middle, where the gradient of flow around the slope is low and transport
capacity is small. Transport capacity increases rapidly, however, when slope
steepens. Deposition in row middles consequently decreases rapidly as the row
deviates slightly from the contour. Use figure 11-24 to adjust nonlinearly the
contouring factor where rows are neither on the contour nor directly uphill and
downhill. This adjusted factor also should be weighted for the watershed area,
as well .

239

1.0

0.8

g 0.6

0.2

P = P + a{ 1.0- P )

CONTOUR CONTOUR

HALF OFF
DEGREE OFF CONTOUR

UP & DOWN
HILL

Figure 11-24. — Adjustment factor for being off contour with tillage.

Furrows of some contouring systems lead runoff to one of several flow con-
centrations in small draws. If excess rainfall exceeds storage capacity, the
ridges overtop in the draws. Erosion in these overtopped areas may be analyzed
as concentrated flow erosion. Select parameter values according to instruc-
tions for a concentrated flow element that requires estimating the depth of
soil from the bottom of the channel to the nonerodible layer. Consider the
nonerodible layer to be at the depth of secondary tillage. Add to this depth
one-third of the difference in height from the top of the row ridge to the bot-
tom of the row middles to obtain a total depth. Use this value as the depth of
the soil beside the channel. Assume a naturally eroded channel shape. If this
approach is used, the overland flow land profile is taken along the rows lead-
ing to the flow concentration.

Graded row middles may act as individual channel systems. These may be
analyzed by assuming that the row ridges are the overland flow area. The aver-
age steepness for the row sideslope is used for slope steepness, and slope
length is the distance from the row ridge to the water's edge in the row mid-
dle. The row middle is described with the model's channel element. Be aware
that this technique and the technique for row breakover have not been validated
for the model .

Hydraulic roughness—Cover and roughness on the soil surface slow overland flow
and reduce its transport capacity. The reduction in velocity depends on the
cover material and its density and the degree of surface roughness. Table
11-26 may be used to select Manning's n values, which the model uses to esti-
mate the transport capacity of overland flow. The ratio of n from table 11-26

240

to that assumed for overland flow over bare soil is a key value. The values in
table 11-26 are based on n = 0.01 for overland flow over bare soil. If that
value is increased, the values in table 11-26 should be changed to maintain the
same ratio of n for cover to n for bare soil. Conversely, if the values in
table 11-26 are adjusted as a whole, similarly adjust the n for bare soil.

Table 11-26. Estimates of Manning's n for overland flow and soil coversl/

Manning's
n

Treatment

Manning s
n

Treatment

Cornstalk residue applied to
fallow surface:

1 ton/acre 0.020

2 tons/acre .040

4 tons/acre .070

Cornstalk residue disk-harrow
incorporated:

1 ton/acre 0.012

2 tons/acre .020

4 tons/acre .023

Wheat straw mulch:

0.25 ton/acre - - 0.015

0.5 ton/acre .018

1 ton/acre .032

2 tons/acre .070

4 tons/acre- ------ .074

Crushed stone

mulch

15 tons/acre -

0.012

60 tons/acre -

.023

135 tons/acre -

.046

240 tons/acre -

.074

375 tons/acre -

.074

Small grain

(20% to full maturity)

Across slope

Grass

Sparse 0.015

Poor .023

Fair .032

Good .046

Excellent .074

Dense .150

Very dense ----- .400

Rough surface depressions

4 to 5 in deep- -0.046
2 to 4 in deep- - .023
1 to 2 in deep- - .014
No surface- - - - .010
depressions

Upslope &
Downslope

Poor stand 0.018

Moderate stand .023

Good stand .032

Dense- ------------- .046

■0.012

■ .015

■ .023

■ .032

1/ Based on data form Lane and others (5_) and Neibling and Foster (_7) •

Overland flow transport is related to Manning's n for bare soil, the ratio
of n with bare soil to that with cover or roughness, Yalin's constant, particle
characteristics, and the deposition reaction coefficent. With the exception of
the reaction coefficient, which can be changed only by an internal program
modification, all these variables must be considered during any optimization of
transport capacity.

241

Strip cropping and grass buffer strips reduce yield of sediment because
close growing vegetation slows the runoff, greatly reducing its transport capa-
city to where deposition occurs. If these practices are not on the contour,
causing flow to move along the strips, or if flow concentrations submerge the
grass, their effectiveness is reduced greatly. Flow along the upper edge of a
strip of grass or other dense vegetation may be treated as a naturally eroded
channel with a slope equal to that along the upper edge of the strip. A tri-
angular channel may be assumed to pass concentrated flow through a strip.

If the problem does not lend itself to treating concentrated flow through
the grass as a channel element, decrease Manning's n to account for a lesser
reduction in flow velocity with the deeper flow. The relation of Manning's n
for flow through grass to the product of velocity and hydraulic radius is not
built into the model. Choose a value that best represents the grass and the
given runoff event.

Channel Element

The channel element describes erosion and sediment transport in flow con-
centrations within farm fields. These are not necessarily defined channels un-
less they happen to be grassed waterways, terrace channels, or diversions. On
many farm fields, flow concentrates in natural, small draws. A heavy rain on
freshly prepared seedbed can cause major erosion in these concentrations. The
soil often will erode down to the depth of secondary or primary tillage. Rid-
ges and field borders can cause flow concentrations at the edge of a field.
The channel element that describes these situations is unsuitable for hydraulic
design of terraces or diversions. Channel sides are assumed to be sufficiently
high to contain the flow.

Non-updateable Parameters

Like the overland flow parameters, there are two sets of nonupdateable
parameters: (1) those that remain constant for the simulation period and (2)
those that are updateable on a storm-by-storm basis. If the nonupdateable
parameters change, the simulation is stopped and restarted. The order in which
the following variables are discussed is changed slightly from the input
order.

Channel shape—The user may specify a triangular, rectangular, or naturally
eroded channel. The triangular channel with equal side slopes is used for most
terrace and grass waterway channels although they are usually parabolic. If
the channel is too parabolic for a triangular section, assume a rectangular
channel. Assume a naturally eroded channel when channel dimensions depend
strongly on previous erosion. Hydraulic calculations are completely indepen-
dent of erosion for the triangular channel. However, erosion rate is limited
by a nonerodible layer and extent of previous erosion. For the rectangular
channel, erosion rate is also limited by a nonerodible layer and extent of
previous erosion. If the calculated eroded width exceeds the specified channel
width, channel width is reset to the eroded width. For the naturally eroded
channel, erosion rate and channel dimensions depend on extent of previous
erosion and the existence of a nonerodible layer.

242

The channel element might be applied to small stream channels (less than 5

to 10 ft wide). Refer to volume I, chapter 3, for a description of the method

used to estimate erosion to consider whether this method is satisfactory in re-
lation to other known methods.

The model should not be applied to gully erosion. Many features of gully
erosion, such as headcutting and sloughing of the sidewalls, are not included
in the channel component.

A single typical flow concentration is chosen to represent flow concentra-
tions of a given stream order. The flow concentration analyzed by the model
may not exist, or it may be one of a number in the watershed, for example, one
out of five in a system of terraces. The output sediment concentration from
the typical flow concentration is assumed to equal the average concentration
for all flow concentrations represented by the chosen channel.

Slope of the energy gradeline (friction slope)--Flow in most channels is dyna-
mic and spatially varied. For many field-sized watersheds, dynamic terms in
the flow momentum equation may be dropped (that is, kinematic assumption). At
your option, normal flow may be assumed to set the friction slope equal to the
slope of the channel. The relatively flat 0.1 to 1/2% slope of many terrace
channels may invalidate this kinematic assumption. Roughness from vegetation
or a ridge at the field's edge may cause backwater not considered by the kine-
matic assumption. In terrace channels where the outlet is unrestricted, flow
accelerates near the outlet, producing higher shear stress than the normal flow
assumption would calculate. At the upper end, shear stresses are lower than
those from the normal flow assumptions.

The normal flow assumption will not work for a zero grade terrace channel
unless slope of the energy gradeline is used as slope input. The normal slope
assumption gives no backwater effect for a restricted outlet unless the slopes
are input as slope of the energy gradeline.

The other option is to use the built-in nonuniform flow curves. These
were developed by normalizing the spatially varied flow equation and solving it
for a range of typical parameter values. Regression analyses were used to fit
polynominal curves to the solutions. Although flow is unsteady, the analysis
assumes steady flow and uses the peak runoff rate as a characteristic discharge
for hydraulic computations. Refer to table 11-27 as a guide to when to use the
two options.

Outlet Control --Four types of control that may be specified are: (1) critical
depth, (2) uniform flow in a downstream control channel, (3) the greater of (1)
or (2), and (4) a structure or control having a known rating curve. Use (1)
for terrace, diversion, or other channels when the depth of flow in the outlet
channel has no restricting effect. Use (2) when a reach at the lower end of
the channel sets the depth (for example, a heavy vegetation at the channel out-
let). Use (3) when the model is to choose the greater of (1) or (2). Use (4)
when a control structure (for example, weir, ridge that acts as a weir, or
flume) controls flow depth according to a known rating curve.

The outlet control is used only when the friction slope curves are used.
This control determines depth at the channel outlet, which is a parameter in

243

Table 11-27. — Guidelines on using the built-in friction slope curves

Condition Friction slope assumption

Supercritical flow all along the channel and at Kinematic (normal) flow,

the outlet.

Small discharge; very flat channel gradient Kinematic (normal) flow.

(0.001) to 0.005); critical depth at outlet
(for example, channel flow in a row middle).

Restricted outlet giving backwater. Friction slope curves.

Critical depth at the outlet of diversions Friction slope curves,

and conventional terrace channels.

Zero grade channel (for example, level terraces). Friction slope curves.

the friction slope curves. If (1), (2), or (3) is specified, a triangular
channel section and its sideslope must be selected (ignore any reference to a
rectangular control section). The preferred sideslope is that of the channel
element. If another sideslope is specified, do not use a flatter slope than
that for the channel. When uniform flow control is selected, slope of the out-
let reach and its Manning's n is specified. Refer to a later section for val-
ues of Manning's n. The form of the rating curve is:

Q = RA(Y - YBASE)RB. tU-9l

Specify values for the coefficient RA and exponent RB. The term Y is flow
depth, and YBASE is the minimum depth for flow to begin. Refer to weir and
flume rating tables to determine the parameter values.

Channel length—This is the length of channel from its outlet to its origin.
Channel origin is the point where overland flow has converged to the point that
it can be considered concentrated flow.

Drainage area at the outlet— If the flow concentration being analyzed is the
main stem in the watershed, the drainage area at the outlet is the total area
in the watershed. If the channel is one of several terrace channels, the area
is that drained by the representative channel chosen for analysis. This area
will be smaller than the overland flow area since the channel only drains a
part of the total overland flow area. Figure 11-25 illustrates this area and
the area at the upper end of the channel .

Drainage area at the upper end — If overland flow converges to form the origin
of the channel, drainage area at the upper end equals the overland flow area
draining into the upper end of the channel. This variable is used to compute
discharge at the upper end of the channel. Channels for terrace outlet usually
originate in the middle of the field with a drainage area at the upper end.
Usually, no drainage area is at the upper end of a terrace channel because of
zero discharge at the upper end (fig. 11-25).

244

(NO AREA EXISTS AT
UPPER END OF SECONDARY
FLOW CONCENTRATION)

DRAINAGE AREA AT OUTLET
OF SECONDARY FLOW
CONCENTRATION = AVERAGE
OF SUB-AREAS 1-6

SECONDARY FLOW
CONCENTRATION

MAIN FLOW
CONCENTRATION

SECONDARY AND MAIN FLOW CONCENTRATIONS

AREA AT UPPER
END OF CHANNEL

DRAINAGE AREA AT
CHANNEL OUTLET

MAIN FLOW CONCENTRATION ALONE

Figure 11-25 Channel areas.

Channel section sides1ope--Use a sideslope of 5 for terrace channels and grass
waterways unless more specific information is available. Use 10 or a value in-
dicated by more specific information for concentrated flow in an area regularly
tilled but susceptible to major erosion. Use 20 for flow concentrations caused
by ridges along field boundaries. Even if a rectangular or naturally eroded
channel is specified, approximate and enter the sideslope for the channel.
Sideslope is used by the model to compute the friction slope. An approximate
value for sideslope for a rectangular channel is:

Z = 3 B5/3/Q2/3

[11-10]

where B = bottom width of the rectangular channel and Q = discharge. Use a
weighted discharge based on the square of the discharges for all storms or
equalize the flow area for a depth weighted toward the larger events.

Distance along the channel—For input, x = 0 at the channel outlet and x
increases going upstream in the channel. Internally, the model inverts this
order and reassigns x = 0 to the upper end of the effective channel length.
Effective channel length is defined from the following relationships.

Discharge rate Q-|ow at the lower end of the channel is:

rpAiow

[II-ll]

where ap = characteristic peak excess rainfall rate and A]ow = drainage area
above the lower end of the channel. Discharge rate qup of the upper end of the
channel is:

245

Qup = Vup [H-12]

where Ayp = drainage area above the upper end of the channel. Lateral inflow
rate q-j is:

qi = (Qlow - Qup)Ach [H-13]

where Xch = channel length. Discharge at any point along the channel is:

Q = Qlow " xch qi [11-14]

where xch is the distance upstream from the lower end of the channel. Effec-
tive channel length Xcheff is defined as the xcn where Q = 0, or from the pre-
ceding equation:

xcheff = Qlow/qi » [II-15]

which is the same as:

*cheff = WU-0 " Aup/Alow). [11-16]

Channel slope--Channel slope at a point (that is, at the specified x's rather
than for a segment) is input. These slopes may be estimated from a plot of the
profile of the channel. A minimum of five points is needed to represent a
curved channel profile. Since the model assumes that the channel profile is
curved, abrupt changes in slope can only be approximated. Enter slope values
and locations on either side of the break for best approximations. The loca-
tions can be a minimum of 1 ft apart.

The model defines channel segments equal to 0.1 of the effective length.
If the actual length is short in relation to the effective length, less than
three channel segments may exist. If this occurs, "fool" the model to have at
least three segments (five or more preferred) by specifying a slightly differ-
ent Manning's n (for example, 0.03001 and 0.03000) at x's when segment ends are
desired. This technique also may be used to "fool" the model to obtain a finer
definition of uniform channel segments.

Updateable Channel Parameters

Channel parameters are read with an x where the parameter changes and the
value of the parameter is just upslope of the change. Input x's are specified
as distance upstream from the channel outlet. When the parameters are updated,
the number of changes is read for each parameter along the channel. If a para-
meter does not change from the previous storm, updating of that parameter is
not required. A zero is read for the number of changes along the channel, and
the model uses the parameter value from the last storm.

Changes along a channel—The model can describe changes in all parameters along
the channel, but it primarily represents end or beginning of grass within the
length of the channel. Figure 11-26 shows a typical grassed waterway ending
within a field where the channel flattens to where erosion in the waterway
probably would not occur. The channel from 0 to 100 ft would be tilled. At

246

seedbed time, a sample parameter data set would be:

Tilled portion
Para-
Distance meter
from END value

Manning's n 0.0 0.03

Critical soil shear ------ .0 .10

Critical cover shear- ----- .0 100

Depth to nonerodible- ----- .0 .33

Depth at side of channel- - - - .0 .33

Channel width .0 20

Grassed waterway

Distance
from END

Para-
meter
value

100 0.13

100 .60

100 100

Same for entire channel
length

Refer to following sections for discussion of selection of these parameter
values.

TILLED

100.0

Figure 11-26. — Representation of
channel having a change
cover along channel .

Manning's n--Different covers have
different values for hydraulic rough-
ness, depending on their density,
height, and type of vegetation. Their
hydraulic roughness also depends on
the rigidity of the vegetation and
degree of submergence by the flow.
Although Manning's n may vary over a
considerable range of the product of
velocity and hydraulic radius, the
model assumes that n does not change
with discharge. Table 11-28 gives
estimated n's for several covers.
Chow U ) , Ree and Crow (8) , and hy-
draulics handbooks of USDA's Soil Con-
servation Service provide additional
information. Do not enter a Manning's
n less than the value entered for the
Manning's n for bare soil.

The Manning's n's in table 11-28 are moderate values for a range of the

product, velocity times hydraulic radius. These values generally are for V •

R of 1.0 to 1.5. For high flows that definitely submerge the cover, the n's
are too high.

Refer to the overland flow, Manning's n section,
adjustment of Manning's n from those in table 11-28.

for discussion on

Critical shear stress of the soil—Some soils and soil conditions are more
susceptible to detachment by flow than are others. Although considerable
information exists on critical shear stress, it is contradictory and generally
does not apply to agriculture. For this model, estimate a base critical shear
stress (lb/ft*) modified from Smerdon and Beasley's equation (10):

rcr = 0.213/dr0-63

[II-17]

247

Table 11-28. — Manning's n for typical soil covers

Cover

Cover density

Manning' s

Smooth, bare soil ;

Less than 1 in deep

roughness elements.

1-2 in deep

2-4 in deep

4-6 in deep

Corn stalks (assumes

1 ton/acre

residue stays in place

2 tons/acre

and is not washed away).

3 tons/acre

4 tons/acre

Wheat straw (assumes

1 ton/acre

residue stays in place

1 .5 tons/acre

and is not washed away).

2 tons/acre

4 tons/acre

Grass (assumes grass

Sparse

is erect and as deep

Poor

as the flow) .

Fair

Good

Excellent

Dense

^ery dense

Small grain

Poor, 7 in rows

(20% to full maturity-

Poor, 14 in rows

rows with flow) .

Good, 7 in rows

Good, 14 in rows

(Rows across flows)

Good

Sorghum and cotton

Poor

Good

Sudangrass

Good

Lespedeza

Good

Lovegrass

Good

0.030
.033
.038
.045

.050
.075
.100
.13

.060
.100
.15
.25

.04
.05
.06
.08
.13
.20
.30

.13
.13
.30
.20

.30

.07
.09

.20

.10

.15

radius.

Does not include effects of submergence or product of velocity-hydraulic

where d,

dispersion ratio. Dispersion ratio is the ratio expressed as a

percentage of suspension percentage divided by percentage of silt plus percent-
age of clay. Typical values for dispersion ratio range from 5 to 25 with most
about 10 to 12. The base critical shear stress applies to a finely pulverized
seedbed, which usually occurs when the soil is most susceptible to erosion.

248

A value of 0.05 lb/ft^ may be used for this base value if no better informa-
tion is available. The soil gradually consolidates and becomes less erodible.
Critical shear stress seems to decrease as tillage more finely pulverizes the
soil. Use figure 11-27 to estimate a factor to multiply the base critical
shear stress for an estimate of apparent critical shear stress. Table 11-29
may be used for critical shear stress if differences in types of soil are not
considered.

40.0

I 2 3 45
CLASS FOR TILLAGE
AND TIME SINCE TILLAGE

TILLAGE CLASS FOR TYPICAL MIDWESTERN
SILT LOAM SOIL

1. LONG TERM WITHOUT TILLAGE

2. ONE YEAR SINCE SEEDBED TILLAGE

3. PRIMARY TILLAGE IN LAND ONE YEAR
SINCE SEEDBED

4. TYPICAL SEEDBED

5. FINELY PULVERIZED SEEDBED

ADJUST DOWN FOR SOILS THAT DO NOT
HAVE A TENDENCY FOR SOIL PARTICLES
TO BOND TOGETHER (e.g. SANDS)

Figure 11-27. — Effect of tillage on critical shear stress.

Table 11-29. — Critical shear stress values as a function of tillage and consol-
idation for moderately erodible soils

Ti 1 1 age-consol idati on
condition Critical shear stress

(Ib/ftZ)

Moldboard plowed - 0.20

Chisel or disk for primary tillage ----------------.15

Disking for common seedbed for corn or cultivation of crop - - - - .10

Finely pulverized seedbed- --------------------.05

1 month after last tillage of common seedbed -----------.20

2 months after last tillage of common seedbed- ----------.30

3 months after last tillage of common seedbed- ----------.40

Long term, undisturbed ----------------------.60

When flow bends vegetation over so that it lies flat on the channel, the
vegetation effectively armors the soil and prevents erosion. The model cannot
directly handle this problem. If this effect is suspected, increase the criti-

249

cal shear stress for the soil to prevent erosion, and decrease n to prevent
deposition, which is unlikely with the flattened vegetation.

Shear stress at failure of crop residue—As in conservation tillage, shear
stress for concentrated flow through and over crop residue may exceed a criti-
cal shear stress at which the cover may begin to move. The model assumes that
if shear stress on the cover exceeds a critical shear stress for the particular
type and rate of a mulch, the cover fails and shear stress is computed as if no
cover exists. The failed cover assumption remains in effect until a new set of
Manning's n is read. Estimates for this critical shear stress are in table
11-30. Assign a value of 100.0 to the variable if cover failure is not
allowed.

Table 11-30 Critical shear stress value for corn stalk and wheat stalk en-

masse movementl/

Type of mulch Rate Critical shear stress

(tons/acre) (lb/ft2)
Corn stalks 1 0.051

(not incorporated). 2 .105

3 .156

4 .210

Wheat straw 1 .064

(not anchored) 1.5 .140

2 .232

4 .841

JL/Stress acting on mulch; does not include stress acting on soil.

Depth of soil from channel bottom to nonerodible layer—When concentrated flow
in a farm field erodes, it often erodes through the tilled layer until it
strikes a nontilled layer and then it rapidly widens. The nonerodible layer is
frequently at the bottom of the surface layer of secondary tillage, which typi-
cally is 0.3 to 0.4 ft deep. Although primary tillage disturbs the soil to a
greater depth than secondary tillage, the large soil chunks turned over by pri-
mary tillage are much less erodible than the surface soils that have been expo-
sed to secondary tillage. The large soil chunks may act like grade control
structures.

In a natural channel, a rock layer or an armor layer acts as a nonerodible
layer. Large flows can destroy the armor layer, however, and the channel will
deepen again until a new armor layer develops. A pond or a larger stream also
may control like a nonerodible layer. The model cannot describe, however, the
development of a concave channel profile upslope from a control like a pond.

Depths to the nonerodible layer are shown in figure 11-28. Whenever til-
lage occurs, this value should be reset. If it is not reset, the model uses
the depth that the channel has eroded to during the previous storm. If the
effect of the nonerodible layer is to be neglected, assign a large value, for
example, 1000.0, to this parameter.

250

ORIGINAL CROSS SECTION

DEPTH TO
NONERODIBLE LAYER
AT SIDE OF CHANNEL

NONERODIBLE LAYER

DEPTH TO NONERODIBLE LAYER
IN MIDDLE OF CHANNEL

SMALL INCISED CHANNEL

SOIL SURFACE
BEFORE EROSION

ERODIBLE TILLED ZONE

DEPTH TO
NONERODIBLE LAYER
AT SIDE OF CHANNEL

NONERODIBLE LAYER

DEPTH TO NONERODIBLE LAYER
IN MIDDLE OF CHANNEL

CONCENTRATED FLOW WATERWAY

Figure 11-28. — Defining sketch for depths to nonerodible layer for a small,
encised channel and for a concentrated-flow waterway.

Depth to nonerodible layer at the side of the channel — If the channel is a flow
concentration through the field in a regularly tilled area, use the same value
as the depth to nonerodible layer in the middle of the channel. For more de-
fined, incised channels, use the height of the effective channel wall that
moves horizontally as the channel widens.

Channel width—Specify channel width when a rectangular channel section is as-
sumed. When a triangular or naturally eroded section is analyzed, specify the
width of the rectangular channel that most closely approximates the channel.
Do not leave this parameter blank. In some situations of no erosion, the model
defaults to a rectangular section.

Pond (Impoundment) Element

Relationships for the pond element were derived from analysis of output
from a simulation model (_3) supported with field observations from impoundment
terraces (4) .

The pond element is primarily meant to describe deposition in impoundment
terraces with pipe outlets, which drain between storms. The pond element can
describe deposition in ponds and impounded water behind ridges and culverts,

251

but it should not be applied to those situations unless discharge is controlled
by a pipe outlet or the equivalent orifice coefficient is known and the im-
poundment drains between storms.

The pond element makes now allowance for short circuiting where the sedi-
ment load enters near the outlet or the entrance is connected directly with a
flow path to the outlet. The pond is assumed to drain completely after the
runoff event and to pass all runoff except for infiltrated water. If some
storm water is retained, the effect on sediment yield must be accounted for
outside the model .

Initial Parameters

If any pond parameters change with time, the simulation must be stopped
when the parameters change and restarted with new parameter values.

Control --The user may specify one of two possible controls: pipe outlet-riser
as typical in parallel tile outlet terraces, or control where the equivalent
orifice coefficient is available. Refer to volume I, chapter 3 for a defini-
tion of equivalent orifice. This option may be unreliable unless it is for the
impoundment terrace type of drain.

Surface area-depth--Values are required for the coefficient and exponent for
the surface area-depth relationship given by:

SA = Fs yB [11-18]

where SA = surface area (ft^), Fs = coefficient, B = exponent, and y =
water depth in the pond (ft). The coefficint and exponent depend on topography
within the ponded area. These values sometimes can be determined from design-
construction surveys. Table 11-31 shows typical values for some impoundment
terraces.

Table 11-31. — Coefficient and exponent for surface area-depth relationships
observed for typical impoundment terraces

Terrace location Coefficient Fs Exponent B

Eldora, Iowa^ 8,247 1.10

Charles City, Iowa^ 9,465 1.73

Guthrie Center, Iowa^ 4,485 1.28

Marvyn, Ala.-/ 7,950 1.77

-^rom Laflen (4).

-From Rochester and Busch (10) .

Values for front, draw, and side slopes may be used by the model to esti-
mate Fs and B if values for them are unavailable. These slopes are front

252

(embankment front slope), draw (slope at the pond along draw draining into
pond), and side (slope of land at the pond toward the draw). The exponent B is
assumed to be 2 and coefficent Fs is calculated from {9):

Fs = C(f + d)/f]2/(d • s) [11-19]

where f = front slope, s = side slope, d = draw slope.

Drainage area — This land area drains into the pond. It is generally the water-
shed area since the pond is assumed to be the last element.

Intake rate—Intake rate is the infiltration rate within the pond and not on
the watershed. It depends on type of soil, sealing, and tillage through the
impoundment area. Refer to the soil survey of USDA-SCS for an indication of
permeability with adjustments for sealing and tillage. A typical value for a
silt loam soil with good intake would be 0.4 in/hr.

Diameter of orifice in outlet pipe--An orifice of a small diameter delays pas-
sage of the runoff through the impoundment and increases depositon. Diameter
of the orifice usually is selected based on volume of impoundment, runoff rates
and volumes, and time to drain. Consult designers of these terraces (usually
USDA-SCS) in your local area to determine actual sizes for the given terraces
or an estimate of typical sizes. If no value can be found, use 3.0 in.

Orifice coefficient— The model actually requires an estimate of C in the equa-
tion:

C = 3600 Q/Y1/2 [11-19]

where Q = peak discharge (ft^/s) out of the pond and Y = depth (ft) of water
above control. If a value for this coefficient is known, it may be input
directly into the model. Otherwise, the model estimates it from the diameter
of the pipe orifice.

OUTPUT

The user gets basic output from the model describing basic parameter val-
ues for the watershed (fig. 11-29). The user can select additional output in
various levels of detail. The first option is an annual summary for each year
in the simulation period, giving sediment yield from the most downstream ele-
ment in the sequence. Sediment yield is for all the types of particle and for
each type individually. Totals for the entire simulation period, also are
given. Figure 11-30 illustrates a summary output.

The second option provides monthly and annual summaries, as shown in fig-
ure 11-31.

The third option summarizes information for each storm and for each ele-
ment in addition to sediment yield. Figure 11-32 illustrates this output.

The fourth option is output from a single storm where loss or deposition
of soil is given for each segment in each element. Figure 11-33 illustrates

253

NONPOINT SOURCE POLLUTION MODEL (EROSION/SEDIMENT VIELD)

EROSION PARAMETERS - GEORGIA PIEDMONT

MANAGEMENT PRACTICE ONE
CONTINOUS CORN - CONUENTIONAL TILLAGE

INITIAL CONSTANTS

STARTING DATE FOR THIS RUN 74000

UT. DENSITY SOIL (IN PLACE) SG.O

WT. DENSITY WATER G2.4

MASS DENSITY WATER 1.94

ACC. DUE TO GRAUITY 32. 2
KINEMATIC UISC05ITY 0.121E-04

MANNING N BARE SOIL (OUER) 0.010

MANNING N BARE SOIL (CHAN) 0.030

CHANNEL ERODIBILITY FACTOR 0.135

(LBS/FT**2 SEC)/(LBS/FT

YALIN CONSTANT (ALL PART.) 0.G35
MOMENTUM COEFF. FOR
NONUNIFORM UELOCITY

IN CROSS SECTION 1.5G (NO UNITS)

JULIAN DATE

LBSF/FT**3

SLUGS/FT**3

FT/SEC**2

FT**2/SEC

*2)««1.05

DISTRIBUTION OF PRIMARY FARTICLES
AND ORGANIC MATTER IN THE ORIGINAL SOIL MASS

TYPE

FRACTION

SPECIFIC SURFACE

(M**2/G OF SOIL)

CLAY

0.140

20.000

SILT

0.200

4.000

SAND

O.GGO

0.050

(M**2

'GANIC MATTER

0.010

1000.000

(ORGANIC CARBON = ORGANIC MATTER/1.73)
INDEX OF SPECIFIC SURFACE 9.38 M**2/G OF TOTAL SOIL

Figure 11-29 — Basic input values for the erosion model.

254

PARTICLE SPECIFICATIONS

TYPE

DIA.

EQSAND DIA.

FALL UEL.

SPGRAU.

FRAC. IN

NO.

FT/SEC

GM/CM**3

DETACH. SED,

0.002

0.102E-04

2. GO

0.03

0.010

0.2B3E-03

2.G5

0.03

0.030

0.020

0..1.25E-C2

1.80

0.23

0.280

0.158

0.542E-01

1.G0

0.27

0.200

0.201
PARTICLE

0.759E-01
COMPOSITION

2.G5

0.45

TYPE

PRIMARY

PARTICLE FRACTIONS

NO.

CLAY
1.000

SILT
0.000

SAND
0.000

ORGANIC MATTER

0.071

0.000

1.000

0.000

0.412

0.588

0.000

0.029

0.070

0.153

0 . 777

0.005

0.000

1.00U

0.000

OUERLAND INPUTS

OUERLAND AREA
SLOPE LENGTH
MAXIMUM ELEUATION
AUERAGE SLOPE
SLOPE OF UPPER END
SLOPE OF MID SECTION
SLOPE OF LONER END

3.2000 ACRES
205,00 FT
5.50 FT
0.02G7
0.0200
0.0380
0.0240

THE SLOPE IS A CONUEX CONCAUE

LOCATION OF UNIFORM SECTION

DISTANCE, ELEUATION 98.0, 3.5

DISTANCE, ELEUATION 15S.0* 1.3

DISTANCE MEASURED FROM THE UPPER END

ELEUATION MEASURED ABOUE LOWEST POINT

Figure 11-29. — Basic input values for the erosion model --continued

255

CHANNEL INPUTS

CHANNEL LENGTH 371.00 FT

DRAINAGE AREA UPPER END 0.2000 ACRES

EFFCT. LENGTH UPPER END 24.73 FT

DRAINAGE AREA LOWER END 3.2000 ACRES

EFFCT. LENGTH LOUER END 395.73 FT
MANNING N FOR RARE SOIL 0.030
SOIL ERODIBILITY FACTOR 0.135

A TRIANGULAR SHAPED CHANNEL

ENERGY GRADELINE
USES THE ENERGY GRADELINE CURUES

RATING CURUE CONTROL

0 = RA*(Y-YBASE)*--RN
RA = 2.410
RN = 2.250
YBASE = 0.00

Figure 11-29. — Basic input values for the erosion model --continued

256

ANNUAL SUMMARY FOR 1974

PART.
TYPE

1
2
3

4
5

TOTAL

67 STORMS PRODUCED
13 STORES PRODUCED

40.26 IN. OF RAINFALL
3.49 IN. OF RUNOFF

QUANTITY OF ERODED SEDIMENT

SOIL LOSS
LBS.

1354.

1213.
10337.

801G.
11046.

31966.

CONCENTRATIONS (SOIL/WATER)
LBSF/FT**3 LBSF/LBSF PPM <WT)

0.0334
0.0299
0.2550
0.1978
0.2726

0 . 7887

0.0005
0.0005
0.0041
0.0032
0.0044

0.0126

535.

480.
4087.
3170.
43S8.

12640.

AUERAGE SOIL LOSS FOR AREA 5.00 TONS/ACRE
(AREA = 3.2000 ACRES)

DISTRIBUTION OF PRIMARY PARTICLES
AND ORGANIC MATTER IN THE ERODED SEDIMENT

TYPE

CLAY
SILT
SAND
ORGANIC MATTER

0.193
0.26S
0.540
0.014

INDEX OF SPECIFIC SURFACE 12.86 M**2/'G OF TOTAL SEDIMENT
ENRICHMENT RATIO OF SPECIFIC SURFACE 1.371

Figure 11-30. — Annual and total summaries from the
erosion/sediment yield model .

257

ANNUAL SUMMARY FOR 1975

STORMS

PRODUCED 48,

,25

IN. OF RAINFALL

STORMS

PRODUCED 7,

.49

IN. OF RUNOFF

THE QUANTITY OF ERODED

SEDIMENT

PART.

IL LOSS

CONCENTRATIONS (SOIL/WATER)

TYPE

LBS.

LBSF/FT**3

_BSF/LBSF

PPM (UT)

1G21.

0.0186

0.0003

298.

1454.

0.01G7

C.0003

2S8.

12215.

0.1404

0.0022

2249.

7771.

0.0893

0.0014

1431.

8254.

0.0948

0.0015

1520.

TOTAL

31315.

0.3598

0.0058

57GB.

AUERAGE SOIL LOSS FOR AREA 4,90 TONS/ACRE
(AREA = 3.2000 ACRES)

DISTRIBUTION OF PRIMARY PARTICLES
AND ORGANIC MATTER IN THE ERODED SEDIMENT

TYPE

CLAY
SILT
SAND
ORGANIC MATTER

FRACTION

0.230
0.314
0.45G
0.01G

INDEX OF SPECIFIC SURFACE 15. 2G M**2/G OF TOTAL SEDIMENT
ENRICHMENT RATIO OF SPECIFIC SURFACE 1.G28

igure 11-30 Annual and total summaries from the

erosion/sediment yield model --continued.

258

NONPOINT SOURCE POLLUTION MODEL (EROSION/SEDIMENT YIELD)

EROSION PARAMETERS - GEORGIA PIEDMONT

MANAGEMENT PRACTICE ONE
CONTINOUS CORN - CONUENTIGNAL TILLAGE

STORM SUMMARY

138 STORMS PRODUCED 88.51 IN. OF RAINFALL

33 STORMS PRODUCED 10.38 IN. OF RUNOFF

THE QUANTITY OF ERODED SEDIMENT IN RUNOFF
UALUES FOR ALL STORMS

PART.

SOIL LOSS

CONCENTRATIONS (SOIL/WATER)

TYPE

LBS.

LBSF/FT**3

LBSF/LBSF

PPM (UT)

2375.

0.0233

0.0004

374.

2GG8.

0.0203

0.0003

335.

22552.

0.17G8

0.0028

2833.

15787.

0.1230

0.0020

1983.

13300.

0.1513

0.0024

2425.

TOTAL G3281. 0.49G1 0.0080 7950.

TOTAL SOIL LOSS FOR AREA 3.90 TONS/ACRE
(AREA = 3.2000 ACRES)

DISTRIBUTION OF PRIMARY PARTICLES
AND ORGANIC MATTER IN THE ERODED SEDIMENT

TYPE FRACTION

CLAY 0.211

SILT 0.290

SAND 0.433

ORGANIC MATTER 0.015

INDEX OF SPECIFIC SURFACE 14.05 M**2/G OF TOTAL SEDIMENT

ENRICHMENT RATIO OF SPECIFIC SURFACE 1.498

Figure 11-30. — Annual and total summaries from the
erosion/sediment yield model --continued.

259

MONTHLY SUMMARY FOR MAY, 1374

10 STORMS PRODUCED
1 STORMS FRODUCED

5.42 IN. OF RAINFALL
0.G4 IN. OF RUNOFF

THE QUANTITY OF ERODED SEDIMENT

PART.
TYPE

1
2
3
4
5

TOTAL

SOIL LOSS
LBS.

38G.
357.

3088.
2315.
3283.

3435.

CONCENTRATIONS (SOIL/WATER)
LBSF/FT**3 LBSF/LBSF PPM CWT)

0.0520
0.0482
0.41G3
0.3125
0.4433

1.273G

0.0008
0.0008
0.00G7
0.0050
0.0071

0.0204

834.

773.
GG80.
5008.
7114.

20410.

AUERAGE SOIL LOSS FOR AREA 1.48 TONS/ACRE
(AREA = 3.2000 ACRES)

DISTRIBUTION OF PRIMARY FARTICLES
AND ORGANIC MATTER IN THE ERODED SEDIMENT

TYPE

CLAY
SILT

SAND
ORGANIC MATTER

FRACTION

0.193
0.2G8
0.533
0.C14

INDEX OF SPECIFIC SURFACE 12.85 M**2/G OF TOTAL SEDIMENT
ENRICHMENT RATIO OF SPECIFIC SURFACE 1.370

Figure 11-31. — Sample of monthly summaries from the
erosion/sediment yield model .

260

MONTHLY SUMMARY FOR JUN, 1974

4 STORMS PRODUCED 5.29 IN. OF RAINFALL
1 STORMS PRODUCED 1.E5 IN. OF RUNOFF

THE QUANTITY OF ERODED SEDIMENT

PART. SOIL LOSS CONCENTRATIONS (SOIL/HATER)
TYPE LBS. LBSF/FT**3 LESF/LBSF PPM (MT)

1 725. 0.0498 0.0008 798.

2 672. 0.04S1 0.0007 739.

3 5827. 0.4000 0.00S4 6411.

4 4979. 0.3418 i",0055 5478.

5 753G. 0.5173 0.0033 8291.

TOTAL 19739. 1.3551 0.0217 2171S.

AUERAGE SOIL LOSS FOR AREA 3.03 TONS/ACRE

(AREA = 3.2000 ACRES)

DISTRIBUTION OF PRIMARY PARTICLES
AND ORGANIC MATTER IN THE ERODED SEDIMENT

TYPE

CLAY
SILT
SAND
ORGANIC MATTER

FRACTION

.176

.246

.578

.013

INDEX OF SPECIFIC SURFACE 11.74 M**2'G OF TOTAL SEDIMENT
ENRICHMENT RATIO OF SPECIFIC SURFACE 1.252

Figure 11-31 Sample of monthly summaries from the

erosion/sediment yield model --continued.

261

THE FOLLOWING PARAMETERS ARE UALID BETWEEN THE DATES (JULIAN)
7400D - 74105

POINTS OF CHANGE ALONG THE OUERLAND PROFILE

DISTANCE

SLOPE

SOIL EROD.

CROPPING

CONTOUR

MANNINGS

FEET

NONDIM.

K FACTOR

C FACTOR

P FACTOR

0.0

0.000

0.020

0.230

0.260

1.000

0.030

33.5

0.454

0.020

0.230

0.260

1.000

0.030

35.0

0.461

0.023

0.230

0.260

1.000

0.030

3S.5

0.4G3

0.023

0.230

0.260

1.000

0.030

38.0

0.47G

0.035

0.230

0«260

1.000

0.030

15G.0

0.757

0.038

0.23C

0 . 260

1.000

0.030

157.4

0.764

0.037

0.230

0.260

1.000

0.030

158.3

0.771

0.036

0.230

0.260

1.000

0.030

1G0.3

0.778

0.034

0.230

0.260

1.000

0.030

1G1.7

0.785

0.033

0.230

0.260

1.000

0.030

163.1

0.732

0.032

0.230

0.260

1.000

0.030

1G4.G

0.733

0.030

0.230

0.260

1.000

0.030

1GG.0

0.806

0.023

0.230

0.2G0

1.000

0.030

1G7.4

0.813

0.027

0.230

0.260

1.000

0.030

168.3

0.820

0.026

0.230

0.260

1.000

0.C30

170.3

0.827

0.025

0.230

0.260

1.000

0.030

206.0

1.000

0.024

0 . 230

0.260

1.000

0.030

POINTS OF CHANGE ALONG THE CHANNEL

STANCE

DISTANCE

SLOPE

MANN. N

WIDTH

DEPTH
MIDDLE

DEPTH
SIDE

SHEAR
CRIT.

SHEAR
COUER

FEET

NONDIM.

FEET

LB/FT**2

24.7

0.063

0.021

0.065

10.000

0.330

0.400

100.000

33.6

0.100

0.021

0.065

10.000

0.330

0.400

100.000

73.1

0.200

0.023

0.065

10.000

0.330

0.230

0.4CO

100.000

118.7

0.300

0.030

0.065

10.000

0.330

0.400

100.000

158.3

0.400

0.027

0.0G5

10.000

0.330

0.400

100.000

137.3

0.500

0.021

0.065

10.000

0.330

0.400

100.000

237.4

0.600

0.015

0.065

10.000

0.330

0.400

100.000

277.0

0.700

0.016

0.065

10.000

0.330

0.400

100.000

316.6

0.800

0.018

0.OC5

10.000

0.330

0.400

100.000

356.2

0.300

0.021

0.065

10.000

0.330

0.400

100.000

335.7

1.000

0.024

0.065

10.000

0.330

0.400

100.000

Figure 11-32. — Sample output from the erosion/sediment yield model showing
storm-by-storm and element-by-element data.

262

STORM INPUTS

DATE 74037 JULIAN DATE

RAINFALL 1.70 INCHES

RUNOFF UOLUME 0.2G INCHES

EXCESS RAINFALL 0.30 INCHES/HR

EI 16.73 UI3CHMEIER ENGL. UNITS

UALUES FOR STORM 74037 FROM OUERLAND FLOW

THE QUANTITY

' OF ERODED

SEDIMENT IN

RUNOFF

PART.

SOIL LOSS

CONCENTRATIONS (SOIL/WATER)

TYPE

LBS.

LBSF/FT**3

LBSF/LBSF

PPM (WT)

34.

0.0112

0.0002

180.

31.

0.0104

0.0002

166.

204.

0.0679

0.0011

1088.

74.

0.0246

C.0004

3S4.

0.0001

0.0000

TOTAL 343. 0.1142 0.0018 1830.
AUERAGE SOIL LOSS FOR AREA 0.05 TONS/ACRE

Figure 11-32. — Sample output from the erosion/sediment yield model showing
storm-by-storm and element-by-element data—continued.

263

UALUES FOR STORM 74037 FROM CHANNEL ONE

PEAK DISCHARGE UPPER END
PEAK DISCHARGE LONER END
CONTROL DEPTH

0.182 FT**3/SEC
2.914 FT**3/SEC
1.088 FT

THE QUANTITY OF ERODED SEDIMENT IN RUNOFF

PART.

SOIL LOSS

CONCENTRATIONS (SOIL/WATER)

TYPE

LBS.

LBSF/FT**3

LBSF/LBSF

PPM (HT)

34.

0.0112

0.0002

180.

31.

0.0103

0.0002

1GG.

201.

0.0G70

0.0011

1073.

43.

0.01G4

0.0003

263.

0.0001

0.0000

TOTAL 315. 0.1051 0.0017 1G84.
AUERAGE SOIL LOSS FOR AREA 0.05 TONS/ACRE

DISTRIBUTION OF PRIMARY FARTICLES
AND ORGANIC MATTER IN THE ERODED SEDIMENT

TYPE

CLAY
SILT
SAND
ORGANIC MATTER

FRACTION

0.380
0.497
0.122

0.027

INDEX OF SPECIFIC SURFACE 25.05 M**2/G OF TOTAL SEDIMENT
ENRICHMENT RATIO OF SPECIFIC SURFACE 2.E72

STORM INPUTS

DATE 74038

JULIAN DATE

RAINFALL 0.20

INCHES

RUNOFF UOLUME 0.00

INCHES

EXCESS RAINFALL 0.00

INCHES/HR

EI O.GG

WISCHMEIER ENGL.

UNITS

*** NO RUNOFF - NO

LOSSES ***

Figure 11-32 Sample output from the erosion/sediment yield model showing

storm-by-storm and element-by-element data—continued.

264

STORM INPUTS

DATE 74178

JULIAN DATE

RAINFALL 4.2t

INCHES

RUNOFF UOLUME 1.E5

INCHES

EXCESS RAINFALL 3.40

INCHES/HR

EI 67.17

UISCHMEIER ENGL,

UNITS

UALUES FOR THE SEGMENT 93.5 FT. FROM THE PROFILE TOP
PARTICLE NET SOIL LOSS RILL SOIL LOSS
TYPE (TONS/ACRE OF SEGMENT)

0.03

0.01

0.03

0.00

0.27

0.04

0.32

0.05

0.55

0.08

TOTAL

UALUES FOR THE SEGMENT 95.0 FT. FROM THE PROFILE TOP
PARTICLE NET SOIL LOSS RILL SOIL LOSS
TYPE (TONS/ACRE OF SEGMENT)

0.04

0.01

0.04

0.01

0.35

0.10

0.41

0.11

0.70

0.20

TOTAL

1.54

0.43

UALUES FOR THE SEGMENT 9G.5 FT. FROM THE PROFILE TOP
PARTICLE NET SOIL LOSS RILL SOIL LOSS
TYPE (TONS/ACRE OF SEGMENT)

0.05

0.02

0.05

0.02

0.44

0.15

0.51

0.17

0.88

0.23

TOTAL

1.93

0.G5

UALUES FOR THE SEGMENT 98.0 FT. FROM THE PROFILE TOP
PARTICLE NET SOIL LOSS RILL SOIL LOSS
TYPE (TONS/ACRE OF SEGMENT)

0.07

0.03

0.0G

0.03

0.5G

0.22

0.G5

0.2G

1.11

0.45

TOTAL

2.45

0.99

Figure 11-33 — Sample output from the erosion/sediment
yield model for successive segments.

265

UALUES FOR THE SEGMENT 156.0 FT. FROM THE PROFILE TOP
PARTICLE NET SOIL LOSS RILL SOIL LOSS
TYPE (TONS/ACRE OF SEGMENT)

0.03

0.05

0.08

0.04

0.73

0.38

0.8G

0.45

1.4G

0.7G

TOTAL 3.23 1.G8

UALUES FOR THE SEGMENT 157.4 FT. FROM THE PROFILE TOP

PARTICLE NET SOIL LOSS RILL SOIL LOSS

TYPE (TONS/ACRE OF SEGMENT)

1 0.10 0.0G

2 0.10 0.0G

3 0.83 0.43

4 0.38 0.57

5 1.G7 0.37

TOTAL 3.S8 2.15

UALUES FOR THE SEGMENT 158.3 FT. FROM THE PROFILE TOP
PARTICLE NET SOIL LOSS RILL SOIL LOSS
TYPE (TONS/ACRE OF SEGMENT)

0.10

0.0G

0.03

0.05

0.80

0.4G

0.34

0.54

1.G0

0.33

TOTAL 3.54 2.05

UALUES FOR THE SEGMENT 1G0.3 FT. FROM THE PROFILE TOP
PARTICLE NET SOIL LOSS RILL SOIL LOSS
TYPE (TONS/ACRE OF SEGMENT)

0.03

0.05

0.03

0.05

0.7G

0.43

0.83

0.51

1.52

0.87

TOTAL 3.3G

UALUES FOR THE SEGMENT 161.7 FT. FROM THE PROFILE TOP
PARTICLE NET SOIL LOSS RILL SOIL LOSS
TYPE (TONS/ACRE OF SEGMENT)

0.03

0.05

0.08

0.05

0.72

0.40

0.84

0.47

1.44

0.80

TOTAL 3.18 1.77

Figure 11-33 — Sample output from the erosion/sediment
yield model for successive segments—continued.

266

UALUES FIT7? THE SEGMENT 1G3.1 FT. FROM THE PROFILE TOP
PARTICLE NET SOIL LOSS RILL SOIL LOSS
TYPE (TONS/ACRE OF SEGMENT)

0.05
0.04
0.37
0.44
0.74

TOTAL 3.00 1.G4

UALUES FOR THE SEGMENT 1G4.G FT. FROM THE PROFILE TOP
PARTICLE NET SOIL LOSS RILL SOIL LOSS
TYPE (TONS/ACRE OF SEGMENT)

0.08

0.G8

0.80

1.3G

0.08

0.04

0.07

0.04

0.G4

0.34

0.75

0.40

1.28

0.G8

TOTAL 2.83 1.51

UALUES FOR THE SEGMENT 1GG.0 FT. FROM THE PROFILE TOP
PARTICLE NET SOIL LOSS RILL SOIL LOSS
TYPE (TONS/ACRE OF SEGMENT)

0.07

0.04

0.07

0.04

0.G0

0.31

0.71

0.37

1.21

0.G3

TOTAL 2.G7 1.39

UALUES FOR THE SEGMENT 1G7.4 FT. FROM THE PROFILE TOP
PARTICLE NET SOIL LOSS RILL SOIL LOSS
TYPE (TONS/ACRE OF SEGMENT)

0.07

0.04

0.07

0.03

0.57

0.23

0.G7

0.34

1.14

0.57

TOTAL 2.51 1.27

Figure 11-33. Sample output from the erosion/sediment

yield model for successive segments — continued.

267

UALUES FOR THE SEGMENT 1G8.9 FT. FROM THE PROFILE TOP
PARTICLE MET SOIL LOSS RILL SOIL LOSS
TYPE (TONS/ACRE OF SEGMENT)

0.07

0.03

0.0G

0.03

0.53

0.2G

0.G2

0.31

LOG

0.52

TOTAL

2.35

1.15

UALUES FOR THE SEGMENT 170.3 FT. FROM THE PROFILE TOP
PARTICLE MET SOIL LOSS RILL SOIL LOSS
TYPE (TONS/ACRE OF SEGMENT)

0.0G

0.03

0.0G

0.03

0.50

0.24

0.58

0.28

1.00

0.47

TOTAL

2.20

1 . 04

UALUES FOR THE SEGMENT 20G.O FT. FROM THE PROFILE TOP
PARTICLE NET SOIL LOSS RILL SOIL LOSS
TYPE (TONS/ACRE OF SEGMENT)

0.0G

0.03

0.0G

0.03

0.49

0.23

0.57

0.27

0.38

0.4G

TOTAL

2.1G

1.02

UALUES FOR STORM 74178 FROM OUERLAND FLOW

THE QUANTITY OF ERODED SEDIMENT IN RUNOFF

PART.

SOIL LOSS

CONCENTRATIONS (SOIL/WATER)

TYPE

LBS.

LBSF/FT**3

LBSF/LBSF

FPM CUT)

372.

0.0255

0.0004

409.

345.

0.0237

0.0004

380.

3013.

0.20G8

0.0033

3315.

3531.

0.2424

0.0039

3885.

G022.

0.4134

0.00GG

GG2G,

TOTAL 13284. 0.9120 0.014G 14615.
AUERAGE SOIL LOSS FOR AREA 2.08 TOMS/ACRE

Figure 11-33 Sample output from the erosion/sediment

yield model for successive segments—continued.

268

UALUES FOR THE SEGMENT 39.6 FT. FROM THE CHANNEL TOP

PARTICLE NET SOIL LOSS CHAN SOIL LOSS

TYPE (LBS/FT OF CHANNEL SEGMENT)

0.08
0.08
0.G7
0.73
1.34

TOTAL 36.52 2.95

UALUES FOR THE SEGMENT 79.1 FT. FROM THE CHANNEL TOP

PARTICLE NET SOIL LOSS CHAN SOIL LOSS

TYPE (LBS/FT OF CHANNEL SEGMENT)

1.02

0.95

8.28

9.71

1G.5G

1.24

0.30

1.15

0.28

10.03

2.41

11.75

2.83

20.04

4.83

TOTAL 44.21 10. G4

UALUES FOR THE SEGMENT 118.7 FT. FROM THE CHANNEL TOP

PARTICLE NET SOIL LOSS CHAN SOIL LOSS

TYPE (LBS/FT OF CHANNEL SEGMENT)

1.80

0.8G

1.G7

0.80

14.57

G.95

17.07

8.15

29.12

13.90

TOTAL G4.22 30. G5

UALUES FOR THE SEGMENT 158.3 FT. FROM THE CHANNEL TOP

PARTICLE NET SOIL LOSS CHAN SOIL LOSS

TYPE (LBS/FT OF CHANNEL SEGMENT)

2. 23

1.35

2.13

1.2G

13.57

10. 96

21.77

12.84

37.13

21.91

TOTAL 81.89 48.32

Figure 11-33. Sample output from the erosion/sediment
yield model for successive segments—continued.

269

UALUES FOR THE SEGMENT 197.3 FT. FROM THE CHANNEL TOP

PARTICLE NET SOIL LOSS CHAN SOIL LOSS

TYPE (LBS/FT OF CHANNEL SEGMENT)

2.31

1.37

a. is

1.27

18. 72

11.11

21.34

13.02

37.42

22.20

TOTAL

UALUES FOR THE SEGMENT 237.4 FT. FROM THE CHANNEL TOP

PARTICLE NET SOIL LOSS CHAN SOIL LOSS

TYPE (LBS/FT OF CHANNEL SEGMENT)

1.33

0.93

1.73

0.92

15. G4

8.03

18.33

9.41

31.27

1G.05

TOTAL G8.97 35.40

UALUES FOR THE SEGMENT 277.0 FT. FROM THE CHANNEL TOP

PARTICLE NET SOIL LOSS CHAN SOIL LOSS

TYPE (LBS/FT OF CHANNEL SEGMENT)

1.80

0.8G

1.G7

0.80

14.57

G.9S

17.08

8.15

23.13

13.91

TOTAL G4.24 30. G7

Figure 11-33. Sample output from the erosion/sediment
yield model for successive segments—continued.

270

UALUES FOR THE SEGMENT 316. S FT. FROM THE CHANNEL TOP

PARTICLE NET SOIL LOSS CHAN SOIL LOSS

TYPE (LBS/FT OF CHANNEL SEGMENT)

1.G5

0.71

1.53

O.GE

13.35

5.74

15. G5

G.73

26. GO

11.48

TOTAL 58.88 25.31

UALUES FOR THE SEGMENT 35G.2 FT. FROM THE CHANNEL TOP

PARTICLE NET SOIL LOSS CHAN SOIL LOSS

TYPE (LBS/FT OF CHANNEL SEGMENT)

2.35

1.41

2.18

1.31

19. OG

11.44

22.34

13.41

38.09

22.88

TOTAL 84.02 50. 48

UALUES FOR THE SEGMENT 395.7 FT. FROM THE CHANNEL TOP

PARTICLE NET SOIL LOSS THAN SOIL LOSS

TYPE (LBS/FT OF CHANNEL SEGMENT)

1.99

1.05

1.81

0.94

14. 8G

7.25

-29.33

-38. 2G

-74.18

-89.40

TOTAL -84.85 -118.42

Figure 11-33. Sample output from the erosion/sediment
yield model for successive segments—continued.

271

UALUES FOR STORM 74178 FROM CHANNEL ONE

PEAK DISCHARGE UPPER END 0.B8G FT**3/SEC
PEAK DISCHARGE LOWER END 10.981 FT**3/SEC
CONTROL DEPTH 1.962 FT

THE QUANTITY OF ERODED SEDIMENT IN RUNOFF

PART.

SOIL LOSS

CONCENTRATIONS (SOIL/WATER)

TYPE

LBS.

LBSF/FT**3

LBSF/LBSF

PPM (W

725.

0.0498

0.0008

798,

G72.

0.04G1

0.0007

739,

5827.

0.4000

0.00G4

6411,

4979.

0.3418

0.0055

5478,

753G.

0.5173

0.0083

8291,

TOTAL 19739. 1.3551 0.0217 21716.
AUERAGE SOIL LOSS FOR AREA 3.09 TONS/ACRE

DISTRIBUTION OF PRIMARY FARTICLES
AND ORGANIC MATTER IN THE ERODED SEDIMENT

TYPE FRACTION

CLAY 0.1 7G

SILT 0.24G

SAND 0.578

ORGANIC MATTER 0.013

INDEX OF SPECIFIC SURFACE 11.74 M**2/G OF TOTAL SEDIMENT

ENRICHMENT RATIO OF SPECIFIC SURFACE 1.252

Figure 11-33. Sample output from the erosion/sediment
yield model for successive segments—continued.

272

this output. The soil loss values for the overland flow or channel segment are
the net loss or gain of sediment from the segment, that is, net loss (or depo-
sition) = [sediment out - sediment in + lateral contribution + flow detachment
(or deposition)]/[area (or length)]. Negative values indicate deposition, but
positive values do not necessarily indicate flow detachment. A positive value
indicates a net loss value, which simply means that more sediment left the
lower end of the segment than entered the upper end. This increase could be
from lateral inflow or lateral inflow plus flow detachment. An increase in the
soil loss per unit watershed area from the overland flow element to the channel
element indicates net channel erosion, while a decrease indicates net channel
deposition.

The option chosen depends on the type of information needed to reach a
management decision. If long-term averages are important, annual summaries are
adequate. If storm-to-storm variability is needed, the third option is chosen.
The fourth option is selected to identify critical areas in the watershed where
rates of erosion or deposition are large and when sediment yield for a design
storm is needed.

MODEL APPLICATION

The intended application of the model is to evaluate sediment yield and
the particle composition of the sediment as influenced by rainfall and runoff,
soil, topography, and management practices. For a given site, management prac-
tices would be studied to identify management schemes that limit total sediment
yield and yield of clay to some tolerable level. Table 11-32 summarizes values
of sediment yield abstracted from simulation runs for 14 storms on 17 different
management practices.

In some situations, makeup of the sediment is as important as the total
amount. Figure 11-30 shows how some situations affect the particle fractions
in the sediment yield.

Interpretation of the results indicates several important considerations.
If the tolerable sediment yield for the simulation time period is 3 tons/acre,
nine practices would be acceptable. Concave slopes, especially those less than
0.5% over an extended distance at the toe, significantly reduce sediment yield
by inducing deposition. Sediment yield from terraces depends on their grade.
Erosion was calculated in the 1% terrace grade, while deposition was calculated
in the 0.25% grade. All terraces are not equally effective in controlling
sediment yield.

Table 11-32 shows that delivery ratio is not constant for all storms and a
single value, such as terraces, cannot be used for a management system. The
delivery ratios in table 11-32 are from model output. The model does not use
delivery ratio to compute sediment yield.

While deposition reduces sediment yield, it segregates the sediment en-
riching the fines (fig. 11-30). Since the composition depends on rainfall and
runoff characteristics, a single design storm is inadequate to evaluate the
effectiveness of best management practices to control pollution.

273

The breadth of the conditions in table 11-32 indicates the ability of the
model to consider such watershed conditions as slope shape, restricted outlets,
eroded drainageways, and a broad range of management practices. Model
parameter values are readily available without calibration. Accuracy of the
results is believed to equal or exceed that of most available models.

Tabl

e 11-32 Typical

model

best management practices that can be analyzed wi
and typical estimates for sediment yield

th the

Practice

All 14

storms

Small

EI =

Runoff =

Sediment
yield

storm

3.6,

0.11 in
Computed
delivery

ratio

Large

EI =

Runoff =

Sediment
yield

i storm

■ 45.4,

■ 1.74 in

Sediment
yield

Computed

delivery

ratio

Computed

delivery

ratio

(tons/acre)

(tons/acre

Conventional

13.18

-71.00

0.21

-71.00

10.63

-71.00

Conventional ,
complex slope
with concave
at toe.

2.29

I/.17

.01

I/.05

2.12

y. 20

Strip cropping,
grass buffer
strip.

.78

I/.06

.00

i/.oo

.72

I/.07

Conventional ,
concentrated
flow.

16.33

i/l.24

.21

-71.00

12.68

i/l.19

Conventional ,
concentrated
flow, restric-
ted outlet.

12.42

I/.94

.14

iy.67

10.04

I/.94

Conventional ,
grass water-
way.

5.40

I/.41

.05

1/.24

4.70

1/.44

Conventional ,
40 ft terrace
interval , 1%
grade.

9.86

1/.75

.10

I/.48

8.88

I/.84

Conventional ,
40 ft terrace
interval 0.8%
grade.

7.00

I/.53

.10

1/.48

6.11

i/.57

Conventional ,
40 ft terrace
interval , 0.5%
grade.

4.62

I/.35

.10

I/.48

3.78

I/.36

274

Table 11-32. — Typical best management practices that can be analyzed with the
model and typical estimates for sediment yield—continued.

Practice

All 14

Sediment
yield

storms
Computed
delivery
ratio

Small

EI =

Runoff =

Sediment
yield

storm
' 3.6,
' 0.11 in
Computed
delivery

ratio

Large

EI ■

Runoff =

Sediment
yield

s storm
' 45.4,
» 1.74 in
Computed
delivery
ratio

(tons/acre)

10.

Conventional ,
40 ft terrace
interval , 0.25%
grade.

2.86

-70.22

0.05

-70.24

2.36

-70.22

11.

Conventional ,
impoundment.

.20

1A.0.2

.01

i/.03

.13

i/.oi

12.

Chisel, 4500
lb/acre, 50%
cover.

2.33

2/ .oo

.02

^l.OO

2.16

2/ 1.00

13.

Chisel, 2000
lb/acre, 20%
cover.

5.87

2/ 1.00

.07

2-/ 1.00

5.31

2/1.00

14.

No-till, 4500
lb/acre, 80%
cover.

.92

2/l.00

.01

2/i.oo

.83

2/1.00

15.

No-till in
killed sod.

.15

2/1.00

.00

2/i.oo

.14

2/1.00

16.

Chisel, 2000
lb/acre, 20%
cover, 40 ft
terrace, 0.5%
grade.

2.41

2/ .41

.01

3/. 09

2.17

-3-/.41

17.

No-till, 4500
lb/acre, 80%
cover, 40 ft
terrace, 0.5%
grade.

1.30

3/1.41

.00

2/ .08

1.22

3/1.47

I/Ratio of sediment yield at outlet to sediment yield from uniform slope,
conventional management.

.2/Ratio of sediment yield with practice to same practice on uniform
slope.

2/Ratio of sediment yield at terrace outlet to sediment yield from uniform
slope with no terraces. Slope length and steepness = 160 ft and 6 pet, respec-
tively. Corn at seedbed time.

275

Table 11-33,

Practice

malysis of several best management practices for the sample
Piedmont watershed

Sediment yield

(SY!

Enrichment ratio (ER)
for specific surface

area

Product
SY*ER

1. Continous corn,

moldboard plow,
disk, cultivate,
unprotected water-
way.

2. Same as (1) ex-

cept grassed
waterway.

3. Same as (1) ex-

cept chisel plow,
no cultivation,
and a grassed
waterway.

4. Same as (1) except

conventional ter-
races on a 0.2%
grade and a grass
outlet channel .

5. Same as (1) ex-

cept impoundment
at lower end of
waterway.

(tons/acre)
6.9

2.4

1.2

1.7

1.8

2.7

2.3

2.8

4.2

12.4

6.5

2. J

4.8

2.9

1/Total for approximately 1-2/3 yr of record.

Practice (1) for the Piedmont watershed reflects sediment yield from the
field where the waterway outlet is restricted, causing ponding and deposition.
Sediment yield from overland flow was estimated at 8.1 tons/acre. This erosion
is calculated but is not printed out when channel elements are used. It was
obtained by rerunning the model and deleting the channel component.

Installing a grass waterway reduces sediment yield by 65%. Although some
of this reduction is due to elimination of erosion in the waterway, much of the
reduction is due to deposition in the waterway as shown by the increased en-
richment ratio. Deposition in the waterway, however, may cause difficult main-
tenance problems.

Chisel plowing limited sediment yield by reducing erosion on the field
surface. Terraces and the impoundment control sediment yield by inducing depo-
sition. Practices that reduce sediment by deposition increase enrichment due
to an increase in the fractions for fines and organic matter.

276

Table 11-34. — Analysis of 3 best management practices for the sample Delta

watershed

Enrichment ratio (ER)
Practice Sediment yield for specific surface Product
(SY)-7 area SY*ER

(tons/acre)

1. Continuous cotton, 15.8 2.5 39.5

fall tillage, mul-
tiple spring til-
lage, grassed field
ditch.

2. Same as (1) except 9.8 2.7 26.5

no fall tillage,
winter cover, and
a 20-ft grassed
buffer strip along
edge of field.

3. Same as (2) except 8.7 2.8 24.4

limited spring til-
lage.

I/Total for 3 yr of record.

The grass buffer strip cut sediment yield by about one-third for the Delta
watershed. Winter cover and reduced tillage somewhat reduced sediment yield.
Even with these practices, the soil is relatively bare for a significant
portion of the year. The enrichment ratios are high because much deposition
occurs in the row middles which significantly enriches the clays. Practically
no sand leaves the field even for the poorest protection.

Grassed waterways significantly reduced sediment yield for the West
Tennessee watershed as well as the Piedmont watershed.

To establish the contribution of overland flow, runs were made with an
overland element alone for both the complex slope shape and a uniform shape
having the same slope as the average slope of the complex slope. Sediment
yield was 17.8 tons/acre from the complex overland flow profile and 76.7
tons/acre from the uniform slope. The difference between 76.7 and 17.8 is due
to slope shape. Much of the difference was due to deposition on the concave
portion on the lower part of the complex slope. The difference between 17.8
and 34.3 tons/acre in table 11-35 for practice 1 is due to erosion by
concentrated flow. Even though grassed waterways controlled erosion by
concentrated flow, erosion on the steeper portions of the overland flow slope
was excessive. A practice such as practice 4 in table 11-35 which controls
both sheet and rill erosion, erosion by concentrated flow, and sediment yield
is desireable.

The effect of most practices on the west Tennessee watershed was similar
to that for the Piedmont watershed. As expected, the enrichment ratio gener-
ally increased as sediment yield decreased. Scatter is great in the relation-

277

Table 11-35. — Analysis of several best management practices for the sample West

Tennessee watershed

Practice

isyj:

Enrichment ratio (ER)
for specific surface Product
area SY*ER

1. Continuous corn,

moldboard plow,
disk, cultivate,
unprotected water-
ways.

2. Same as (1) except

grassed waterways.

(tons/acre]
34.3

12.9

1.9

2.8

65.2

36.1

3. Permanent pasture. .12

4. Same as (1) except 4.9

no-till (3400 lb/acre
60% cover) , grassed
waterway.

5. Same as (1) except 6.5

impoundments at end
of tributary water-
ways.

2.2
2.7

4.2

.2
13.2

27.3

- Total for 3 yr of record,

ship, however. Note the low enrichment ratio for the pasture practice on the
west Tennessee watershed. For this situation, sediment yield was controlled by
detachment, which was limited by surface cover. Conversely, note the large en-
richment ratios for large sediment yield rates for the Delta watershed. These
enrichment ratios are large because deposition controlled sediment yield.
Using this type of model is advantageous in that the model can represent com-
plex interactions.

The product of sediment yield and enrichment ratio is a pollution index
for the sediment in that it measures the amount and fineness of sediment.
Viewed in that perspective of the cropping system analyzed, the best are the
chisel system on the Piedmont watershed, the limited tillage and grass buffer
strip system on the Delta watershed, and the no-till system on the West
Tennessee watershed. Depending on the selected tolerance level, one or more
practices might be acceptable. Of practices that give sediment loads meeting
the tolerance level, the farmer selects the one that best fits his total farm-
ing operation.

278

CALIBRATION

Obviously, model results can be improved by calibration. However, if
calibration is used, the following cautions should be observed.

Carefully inspect the quality of the observed data. Especially make sure
that deposition at the flume did not reduce the data to a measure of transport
capacity through the flume.

Keep the calibrated parameters minimal. On areas of overland flow, the
parameters most likely to be in error are soil erodibility factor (up to a
factor of 2 to 3) and Manning's n (a factor of 2 to 3). The soil loss ratios
represent well the influence of management practices, although in a given situ-
ation they might be off by a factor of 2. For overland flow, therefore, the
calibrated variables should be limited to soil erodibility and Manning's n.

The main calibration factors in the channels, are soil erodibility (off by
a factor of 2 to perhaps 5), critical shear stress (off by a factor of 2 to
perhaps 5), outlet control characteristics, and Manning's n. Manning's n is
reasonably well defined for channels, but the model does not allow it to vary
with discharge or flow depth. If the outlet control is calibrated, use a rat-
ing-curve control .

Although a representative particle size could be selected by calibration,
input the primary particle size and use the default distribution. Distribution
of primary particles of the soil does not represent the distribution of sedi-
ment particles for most agricultural soils.

Intake rate, shape parameters, and orifice coefficients are calibratable
parameters for the pond.

Overall, peak runoff rate should be considered to be a calibratable para-
meter. The user should recognize the magnitude of errors likely in estimating
peak runoff rate. Also, variation of Manning's n values, discussed above,
could affect peak runoff rates.

Use calibration sparingly. Calibration for one management practice will
not insure an adequate evaluation of an alternate management practice on the
same watershed. The model parameter values have been given to minimize the
need for calibration.

279

REFERENCES

(1) Chow, V. T.

1959. Open-channel hydraulics. McGraw-Hill Book Company, Inc., New
York, N.Y., 680 pp.

(2) Davis, S. S.

1978. Deposition of nonuniform sediment by overland flow on concave
slopes. MS Thesis, Purdue University, West Lafayette, Ind., 137 pp.

(3) Laflen, J. M., H. P. Johnson, and R. 0. Hartwig.

1978. Sedimentation modeling of impoundment terraces. Transactions
of the American Society of Agricultural Engineers 21(6) : 1131-1135 .

(4) s h. P. Johnson, and R. C. Reeve.

1972. Soil loss from tile outlet terraces. Journal Soil and Water
Conservation 27(2) :74-77.

(5) Lane, L. J., D. A. Woolhiser, and V. Yevjevich.

1975. Influence of simplification in watershed geometry in simulation
of surface runoff. Hydrology Paper No. 81, Colorado State Univer-
sity, Fort Collins, Colo., 50 pp.

(6) Lombardi, F.

1979. Universal Soil Loss Equation (USLE), runoff erosivity factor,
slope length exponent, and slope steepness exponent for individual
storms. PhD Thesis, Purdue University, West Lafayette, Ind.

(7) Neibling, W. H., and G. R. Foster.

1977. Estimating deposition and sediment yield from overland flow pro-
cesses, jji: D. T. Kao (ed.), Proceedings of the International Sym-
posium on Urban Hydrology, Hydraulics and Sediment Control. UKY-
BU114. University of Kentucky, Lexington, Ken., pp 75-86.

(8) Ree, W. 0., and F. R. Crow.

1977. Friction factors for vegetated waterways of small slope. U. S.
Department of Agriculture, Agricultural Research Service, Southern
Region, ARS-S-151, 56 pp. (Series discontinued; Agricultural Research
Service is now Science and Education Adminstration-Agri cultural
Research. )

(9) Rochester, E. W., and C. D. Busch.

1974. Hydraulic design for impoundment terraces. Transactions of the
American Society of Agricultural Engineers 17(4) :694-696, 700.

(10) Smerdon, E. T., and R. P. Beasley.

1959. Tractive force theory applied to stability of open channels in
cohesive soils. Agricultural Experiment Station, University of
Missouri, Research Bulletin No. 715. Columbia, Mo., 36 pp.

280

(11) Williams, J. R., and H. D. Berndt.

1977. Determining the Universal Soil Loss Equation's length-slope
factor for watersheds. Jhn: Soil Erosion: Prediction and Control.
Soil Conservation Society of America, Ankeny, Iowa. Special
Publication No. 21, pp. 217-225.

(12) Wischmeier, W. H., and D. D. Smith.

1978. Predicting rainfall erosion losses. U.
Agriculture, Agriculture Handbook No. 537, 58 pp.

S. Deparment of

(13) , C. B. Johnson, and B. V. Cross.

1971. A soil erodibility nomograph for farmland and construction
sites. Journal of Soil and Water Conservation 26(5) : 189-193.

14) Yalin, Y. S.

1963. An expression for bed-load transportation. Journal of the
Hydraulics Division, Proceedings of the American Society of Civil
Engineers 89(HY3) :221-250.

281

Chapter 3. NUTRIENT SUBMODEL
M. H. Frere and J. D. Nowlin-^'

INTRODUCTION

This nutrient model was developed to provide the user with estimates of
nitrogen and phosphorus losses from fields. With the model, the user can simu-
late the effects of such best management practices as erosion control practices
or timing and method of nutrient applications. The results of these simula-
tions can be analyzed to determine if any proposed practice increases losses or
which practices most effectively control nutrient losses.

The model was developed with a minimum amount of information needed for a
reasonable or acceptable prediction. Most of the relations used, therefore,
are simple and do not require many parameters which are frequently unavailable.
The simple relations are solved sequentially rather than simultaneously, which
reduces computer time.

Since the variability of physical and chemical parameters across the field
is often + 20%, our goal for overall accuracy was + 40% over several years when
average measured parameter values are used. Since the error in predicting in-
dividual storm events can be considerably greater, a wide variety of climatic
conditions (10-30 yr) , should be used to generate information for probability-
type analysis.

Submodel Structure

A main program calls both pesticide and nutrient subprograms. The six nu-
trient subprograms are NUTRIN, NUT208, NUTEND, NUTRES, NUTANN, and NUTMON. Two
other subroutines, ANNPCP and THEEND, are called to print annual and end-of-
record headings with rainfall and runoff summaries.

Subroutine NUTRIN is called to read in values of the parameters and ini-
tial conditions for operating the model. This subroutine also prints these
values at the beginning of the simulation results to document the values used.

Subroutine NUT208 is the main subprogram that calculates the movement of
nutrients between compartments and subsequent losses of nutrients. The first
section under "initial conditions" establishes the initial conditions for sev-
eral variables and calculates the value of some parameters. These calculations

1/ Soil scientist, USDA-SEA-AR, Southern Region Office, New Orleans, La.,
and computer programmer, Agricultural Engineering Department, Purdue Universi-
ty, West Lafayette, Ind., respectively.

282

are bypassed, except when new conditions are introduced, NEWNT > 0. The rest
of the section calculates the transpiration ratio for the period since the last
storm, TR, the amount of nitrogen added in the current rainfall, RN, and up-
dates or initializes some variables.

When percolation occurred during a period, denitrification (DNI) and ni-
trate leaching (NL) are assumed to have occurred. These values cannot exceed
the current value of nitrate in the root zone, N03. Burns' estimation of
leaching (BNL) also is calculated although it only has significance at the end
of the year.

Following leaching, nutrients from fertilizer, wastes, and residues are
added to the soluble N and P compartments, SOLN and SOLP, and the soil nitrate,
N03, on the day of fertilization, DATE F or DF (ID). If loss of sediment oc-
curred, S0L0SS > 0, N and P losses with the sediment, SEDN and SEDP, are calcu-
lated.

The fraction of rainfall that does not appear as runoff infiltrates the
soil and leaches some N and P out of the surface layer. The leached nitrate is
added to the nitrate in the root zone, which is subject to further changes.
Because of the buffering capacity of the soil, the phosphate level is not per-
mitted to be reduced below the initial soil level.

When there is runoff, RUNOFF > 0, some N and P are lost in the runoff, RON
and ROP, in proportion to the concentration in the surface layer, CN and CP.
These runoff losses cannot exceed the amount available in the surface layer.

Under "mineralization", the average temperature, ATP, is used to modify
the rate constant, TK. The amount mineralized, MN, is calculated from the
amount of potentially mineral izable nitrogen, POTM, and the soil water correc-
tion, WK, equal to the ratio of average water content to field capacity.

Under "uptake", the plant only takes up nitrogen from the nitrate in the

soil between the date of emergence, DEMERG, and the harvest date, DHRVST . The

value of OPT determines which option will be used for calculating nitrogen up-
take.

In option 1, the amount of dry matter, DM, is calculated from the yield
potential, YP, and the ratio of actual transpiration, ACTSP, to potential water
use, PWU. The fraction of the total growth expected is calculated using the
fraction of remaining potential transpiration, 1-SWU/PWU. The concentration of
nitrogen in the plant material changes with growth and is the minimum value
calculated from two power equations. The amount of nitrogen currently in the
plant is the product of the dry matter and the concentration in the dry matter.
Uptake, UP, is the difference between the current and the previous value.

In option 2, the time since emergence, T, is used to compute normalized
probability variate, X, from the mean, DOM, and standard deviation, SD, both in
days. The fraction of potential uptake is calculated using a fourth order
polynominal representation of the probability curve. The actual nitrogen in
the plant material is calculated using the amount of potential uptake, PU, and
the transpiration ratio, TR, to account for water stress.

283

The amount of uptake from either option cannot be higher than the amount
of nitrate in the root zone, N03. Finally, the total runoff and sediment loss-
es of N and P are accumulated.

The subroutine NUTRES prints out the losses of N and P after each storm.
If no runoff occurs, only the uptake, mineralization, and drainage losses are
printed. The average concentration of N and P in the runoff waters is PPMN and
PPMP.

This information is useful in identifying when and, hence, under what con-
ditions, the highest losses occur. It can be used to select management practi-
ces that might be effective. Since some storms are more frequent than others,
the information printed out by the subroutine can be used to develop probabili-
ty of occurrence graphs. These graphs are most useful in comparing the effects
of management practices.

Subroutine NUTMON is called at the end of each month to print out monthly
summaries of nutrient losses. The monthly summaries provide a convenient means
of reviewing losses on a seasonal basis relative to rainfall and management re-
gimes.

The subroutine ANNNUT is called at the end of the year to print out the
accumulated nutrient losses during that year. This annual summary is useful
because nutrient problems tend to be chronic rather than acute. Therefore, an-
nual loads leaving a field probably are a better reflection of impact than any
single storm event.

The subroutine NUTEND is called at the end of the simulation period to
print out the total accumulated nutrient losses. This final summary is the
best single value for evaluating best management practices for controlling nu-
trient pollution. Since nutrient problems are long term, the losses accumula-
ted over many storms and years better reflect the average effects.

SELECTION OF VALUES FOR INPUT DATA

Storm/Hydrology/Erosion Data File

This file is created in the hydrology and erosion components of the model
passed from the erosion component. Table 11-36 shows the card format, variable
name, and variable definition for the data from the erosion pass file. A sam-
ple card image arrangement for the pass file is shown in figure 11-34. This
file would not need to be recreated if various fertilization practices were
evaluated or if the nutrient model itself was evaluated.

SDATE is the Julian, date of the storm, including both the last two digits
of the year and the day number, for example, 74123.

RNFALL is the inches of rainfall occurring in a storm on that date. It is
converted to millimeters for use in the nutrient model.

RUNOFF is the inches of runoff from the storm and is converted to milli-
meters for the nutrient model.

284

Table 11-36. — Chemistry model input

Storm/Hydrology/Erosion Data File

Card 1. SDATE, RNFALL, RUNOFF, SOLQSS, ENRICH, DP, PERCOL, AVGTMP,

AVGSWC, ACCPEV, POTPEV, ACCSEV, POTSEV

SDATE Date of storm (Julian date), e.g. 73148

RNFALL Volume rainfall (in|cm), e.g. 4.27|10.85

RUNOFF Volume of runoff (in|cm), e.g. 1.5814.01

SOLOSS Amount of eroded sediment ( tons/ acre! kg/ha) , e.g.
4.3219674.0

ENRICH The sediment enrichment ratio computed with particle
size distribution information, e.g. 1.30

DP Number of days since the last storm when percolation

occured, e.g. 1

PERCOL Percolation below the root zone (in I cm), e.g.
1.01512.58

AVGTMP Average temperature between storms (Degrees F. |C),
e.g. 72.8122.7

AVGSWC Average soil water between storms (in/in) , e.g. 0.3239

ACCPEV Actual EP (evaporation from plants) for the period
between storms (in|cm), e.g. 0.022|0.056

POTPEV Potential EP for the period between storms (in|cm),
e.g. 0.02210.056

ACCSEV Actual ES (evaporation from soil) for the period
between storms (in|cm), e.g. 0.000|0.000

POTSEV Potential ES for the period between storms (in|cm),
e.g. 0.00010.000

285

Table 11-36.—- Chemistry model input—continued

Card 1 is repeated for each rainfall event. Ihe last card on the file
should be blank to indicate the end of data. The Erosion program creates a
file called "SEDPAS" specifically for use as this file. The values in the
Storm/Hydrology/Erosion file are in English units when it is created with the
Erosion program. If the file isn't created with the Erosion program then
either English or Metric units may be used.

A small sample of a typical Storm/Hydrology/Erosion Data file follows.
It will illustrate the file structure.

Format(l6,F6.2,F6.2,F6.2,F6.2,I2,F6.2,F6.2,F6.4,F6.3,F6.3,F6.3,F6.3)

EROSIOM PASS FILE EXAMPLE

0.00 70.730.3171 0.002 0.002 0.000 0.000

0.00 71.530.3138 0.010 0.010 0.000 0.000

0.00 72.180.3191 0.004 0.004 0.000 0.000

1.01 72.810.3239 0.022 0.022 0.000 0.000

0.00 74.230.3509 0.0G5 0.0G5 0.000 0.000

0.38 75.220.3GG4 0.008 0.008 0.000 0.000

0.13 75.530.3S25 0.017 0.017 0.000 0.000

0.20 75.830.3GG4 0.009 0.009 0.000 0.000

731G1 0.25 0.00 0.00 0.00 0 0.00 7G.030.3G54 0.009 0.009 0.000 0.000

731G4 0.78 0.03 0.03 2.91 1 0.19 7G. 400. 3597 0.02G 0.02G 0.000 0.000

73139

0.48

0.00

.00

0.00

73143

0.52

0.00

,00

0.00

73144

0.23

0.00

,00

0.00

73148

4.27

1.58

.34

1.31

7315G

0.28

0.00

,00

0.00

73157

1.22

0.12

, 1 1

2.G3

73159

0.G0

0.01

.01

4.15

731G0

0.50

0.03

.02

2.77

286

(BLANK CARD FLAGS THE END OF THE FILE)

REMAINDER OF THE
STORM/HYDROLOGY/EROSION DATA
(I CARD/EVENT)

FORMAT(I6,F6,2,F6,2,F6,2,F6,2,I2,F«,2,F6,2,F6,«,F6,3,F6,3,F«,3,F6,3)
SDATE RNFALL RUNOFF SOLOSS ENRICH DP PERCOL AVGTMP AVGSWC ACCPEV POTPEV ACCSEV POTSEV

Figure 11-34. — Schematic representation of a sample card deck arrangement and
format for the erosion/sediment yield pass file.

SOLOSS is the tons/acre of soil loss in that storm. It is converted to
kilograms/hectare and called SED in the nutrient model.

ENRICH is an enrichment factor computed in the erosion model but not used
currently in the nutrient model.

DP is the number of days since the last storm when percolation occurred.

PERCOL is the inches of percolation below the root zone since the last
storm. It is converted to millimeters and called PERC in the nutri-
ent model .

AVGTMP is the average Farenheit temperature between storms. It is conver-
ted to Celsius and called ATP in the nutrient model. The tempera-
ture of the soil in the top 2 ft is preferred, but air temperature
is used as an approximation.

AVGSWC is the average volumetric soil water content of the root zone be-
tween storms and is called AWC in the nutrient model.

ACCPEV is the inches of actual transpiration between storms. It is con-
verted to millimeters and called ACTSP in the nutrient model.

POTPEV is the inches of potential transpiration between storms. It is
converted to millimeters and called ACPTSP in the nutrient model.

ACCSEV is the inches of actual soil evaporation between storms. It is
converted to millimeters in the nutrient model.

POTSEV is the inches of potential soil evaporation between the storms and
is converted to millimeters in the nutrient model.

Nutrient Parameter File

The chemistry component of CREAMS contains the plant nutrient submodel as
well as the pesticide submodel. The complete chemistry parameter listing,
given in table 11-37, is used here and again in chapter 4 to prevent confusion

287

Table 11-37. — Chemistry model input parameters file

Initial General Parameter Inputs

Card 1-3.

TITLE ()

TITLE

Card 4.

BDATE,

BDATE

FLGOUT

FLGIN

FLGPST

Three lines of 80 Characters each for alphanumeric
information to be printed at the beginning of the out-
put, format (20A4)

FLGOUT, FLGIN, FLGPST, FLGNUT

The beginning date for simulation. It must be less
than the first storm date (SDATE) . (Julian date), e.g.
73138

0 for annual summary output

1 for annual and monthly summary output

2 for individual storm and all summary output

0 if the Storm Hydrology input is in English units and
will need to be converted to metric

1 if the values are already in metric units

0 if there will be no Pesticide inputs

1 if there will be Pesticide simulation

FLGNUT

0 if there will be no Plant Nutrient inputs

1 if there will be Plant Nutrient simulation

Card 5. SOLPOR, FC, OM

SOLPOR Soil porosity (cc/cc) , e.g. 0.41

FC Field capacity (cc/cc), e.g. 0.32

OM Organic matter available for denitrif ication (% of soil
mass) , e.g. 0.65

Initial Pesticide Inputs

Card 6. NPEST, PBDATE, PEDATE

NPEST Number of pesticides, e.g. 2, MAX of 10

If blank the pesticides portion of the model is
bypassed .

PBDATE Date the model begins to consider pesticides (Julian
date) , e.g. 74120

288

Table 11-37. — Chemistry model input parameters file— continued

PEDATE Date the model stops considering pesticides (Julian
date) , e.g. 75365

Card 7.

Card

Card 9.

Initial Plant Nutrient Inputs
OPT
OPT

1 for option one nitrogen uptake

2 for option two nitrogen uptake

SOLN, SOLP, N03, SOILN, SOILP, EXKN, EXKP, AN, BN, AP

SOLN Soluble nitrogen (kg/ha) , e.g. 0.2

Soluble phosphorous (kg/ha), e.g. 0.2
Nitrate (kg/ha), e.g. 20.0
Soil nitrogen (kg/kg), e.g. 0.00035
Soil phosphorous (kg/kg), e.g. 0.00018
Extraction coefficient for nitrogen, e.g. 0.0576
Extraction coefficient for phosphorous, e.g. 0.07
Enrichment coefficient for nitrogen, e.g. 16.8
Enrichment exponent for nitrogen, e.g. -0.16
Enrichment coefficient for phosphorous, e.g. 11.2

SOLP

NO 3

SOILN

SOILP

EXKN

EXKP

BP, RCN

RCN

Enrichment exponent for phosphorous, e.g. -0.146
Concentration of nitrogen in rainfall (mg/1) , e.g. 0.8

Updateable General Parameter Inputs

The rest of the input to the Chemicals program is updateable. The pro-
gram checks the dates (SDATE, card 1) from the Storm/Hydrology/Erosion file
against the parameters control date (CDATE, card 10) . If the control date is
less than the date of the storm, the program reads in a new set of the update-
able parameters. If the program reads a blank in place of the control date
(CDATE, card 10) the program stops executing.

Card 10.

PDATE, CDATE

289

Table 11-37. — Chemistry model input parameters file—continued

PDATE First date that the following chemical parameters are
valid (Julian), e.g. 73138

The program doesn't read in a value for PDATE. PDATE
is only used as an aid in putting together the data
file.

CDATE Last date that the following chemical parameters are
valid, for example one day before the next pesticide
application or one day before a change in the plant
nutrients parameters (Julian), e.g. 73131

NOTE; A card 10. should always be the first card in a
new set of updateable parameters.

Updateable Pesticide Inputs

Card 11.

APDATE
APDATE

Date the pesticide is applied (Julian date) , e.g. 73121
if blank, cards 12-14 are not read

Card 12.

PSTNAM ( )

PSTNAM

Card 13.

APRATE,

WSHTHR

APRATE

DEPINC

EFFINC

FOLFRC

SOLFRC

FOLRES

The pesticide name, up to 24 characters, format (6A4) ,
e.g. ATRAZINE

APRATE, DEPINC, EFFINC, FOLFRC, SOLFRC, FOLRES, SOLRES, WSHFRC,

Rate of application (kg/ha), e.g. 3.36

Depth of incorporation (cm), e.g. 1.0

Efficiency of incorporation, e.g. 1.0

Fraction of pesticide applied to the foliage, e.g. 0.0

Fraction applied to the soil, e.g. 1.0

Amount of pesticide residue on the foliage prior to
this application (ug/g) , e.g. 0.0

Amount on the soil prior to this application (ug/g) ,
e.g. 0.0

Fraction on the foliage available for rainfall washoff,
e.g. 0.0

Rainfall threshold for foliage washoff (cm), e.g. 0.0

290

SOLRES

WSHFRC

WSHTHR

Table 11-37. — Chemistry model input parameters file—continued

Card 14. SOLH20, HAFLIF, EXTRCT, DECAY, KD

SOLH20 Water solubility (PPM), e.g. 33.0

HAFLIF Foliar residue half life (days), e.g. 0.0

EXTRCT Extraction ratio, e.g. 0.1

DECAY Decay constant, e.g. 0.10

KD KD, e.g. 2.0

Cards 11-14 are repeated for each pesticide (NPEST, card 6). If the
application date (APDATE, card 11) is blank then cards 12-14 are omitted for
that pesticide and the old values, including APDATE, are retained. This is
usefull when one of the pesticides is to be reapplied but others are not. If
more than one pesticide is applied and the pesticides are applied on different
dates, blank cards must be inserted at the appropriate places in the file for
each pesticide not being applied with this update. The following example is
given for clarification.

Assume 3 pesticides (NPEST, card 6) are applied with the following application
dates:

Atrazine - 3/20/74
2,4-D - 4/15/74
Parathion - 6/13/74

6/20/75 (75171)

(74079), 3/27/75 (75086)

(74105) , 4/12/75 (75102)

(74164) , 7/05/74 (74186) ,

7/21/75 (75202)

The following cards 10-14 would be used:

Card 10
Card 11
Card 12
Card 13
2 blank
Card 10
1 blank
Card 11
Card 12
Card 13

1 blank
Card 10

2 blank
Card 11
Card 12
Card 13
Card 10
2 blank
Card 11
Card 12

74000 74104

74079
Atrazine
and 14: appropriate data
card 11' s for 2,4-D and Parathion

74105 74163
card 11 for Atrazine

74105
2,4-D
and 14: appropriate data
card 11 for Parathion

74164 74185
card 11' s for Atrazine and 2,4-D

74164
Parathion
and 14: appropriate data

74186 75085
card 11 's for Atrazine and 2,4-D

74186
Parathion

291

Table 11-37. — Chemistry model input parameters file—continued

Card 13 and 14: appropriate data

Card 10: 75086 75101

Card 11: 75086

Card 12: Atrazine

Card 13 and 14: appropriate data

2 blank card ll's for 2,4-D and Parathion

etc. . .

NOTE: Some computers read a blank card as undefined or some other type of
illegal data that will result in an execution error. A zero punched in the
data fields on blank cards will prevent this from occur ing.

Updateable Plant Nutrient Inputs

Card 15. NF, DEMERG, DHRVST

NF number of fertilizer applications, e.g. 2

DEMFRG Date of plant emergence (Julian date, no year), e.g.
141

DHRVST Date of plant harvesting (Julian date, no year) , e.g.
305

When no new Plant Nutrient values are to be read card 15 should be left
blank. The program will then skip reading the remaining Plant Nutrient param-
eters.

For Option One Nitrogen Uptake

Card 16. RZMAX, YP, DMY, POTM, AWU, PWU

RZMAX Maximum depth of the root zone (mm) , e.g. 450.0

YP Potential yield (kg/ha), e.g. 5700.0

DMY Dry matter yield ratio, e.g. 2.5

POTM Potential mineral izable nitrogen (kg/ha) , e.g. 47.0

AWU Actual water use (mm), e.g. 570.0

PWU Potential water use (mm), e.g. 780.0

Card 17. CI, C2, C3, C4

C1,C3 Cubic coefficients, e.g. 0.0209, 0.0128
C2,C4 Cubic exponents, e.g. -0.157, -0.415

292

Table 11-37 Chemistry model input parameters file—continued

por option Two Nitrogen Uptake

Card 16. RZMAX, YP, DMY, POTM, DOM, SD, PU

RZMAX Maximum depth of the root zone (mm) , e.g. 450.0

YP Potential yield (kg/ha), e.g. 5700.0

DMY Dry matter yield ratio, e.g. 2.5

POTM Potential mineralizable nitrogen (kg/ha) , e.g. 47.0

DOM Date of mid point in nitrogen uptake cycle (days), e.g.
73.0

SD Standard deviation of DOM (days), e.g. 30.0

PU Potential nitrogen uptake (kg/ha) , e.g. 250.0

Both Options Continue

Card 18. DF(1)

DF Date of fertilizer application (Julian date) , e.g.
73131

Card 19. FN(1), FP(1), FA(1)

FN Nitrogen applied (kg/ha), e.g. 28.0

FP Phosphorous applied (kg/ha), e.g. 28.0

FA Surface fraction of application, e.g. 0.1

Cards 17 and 18 are repeated for each application of fertilizer (NF, card
15). A maximum of 20 applications can be read in one update.

293

Table 11-37. — Chemistry model input parameters file — continued

A sample data file for the Control Parameters for the plant nutrients model
follows. It will help demonstrate the file structure.

CARD
NO

2
3
4

9
10
15
1G
13
19
13
19
10
15
1G
18
19
18
19
10
15
1G
18
19
13
19
10
15
1G
10

CHEMISTRY PARAMETER DATA

7313S 1
0.410 0.320

0.200 0.200

-0.14G 0.800

73305

2 141

450.0005700.000

73131

28.000

73174

112.000

28.000

0.000

74305

2 129

450.0005700.000

74119

NUTRIENTS PARAMETERS - GEORGIA PIEDMONT
MANAGEMENT PRACTICE ONE
CONTINOUS CORN - CONUENTIONAL TILLAGE
0 0 1
0.G50

20.000 0.00035 0.00018 0.057G0 0.07000 1G.8000 -0.1G00 11.2000

305
2.500 47.000 73.000 30.000 250.000

0.100

1.000

305
2.500 47.000 73.000 30.000 250.000

28.000

741G2

112.000

28.000

0.000

75305

2 151

450.0005700.000

75141

28.000

7517G

112.000

28.000

0.000

753G5

0 151

450.0005700.000

0.100
1.000

305
2.500 47.000 73.000 30.000 250.000

0.100
1.000

305
2.500 47.000 73.000 30.000 250.000

2 94

of overall program input. Only the nutrient parameter file will be discussed
in this chapter. The chemistry component can be run on the computer with only
nutrient data if the user desires.

General Parameters

Some parameters do not change significantly during a simulation period,
although it is recognized that such changes may be gradual over time. If an
intensive management system significantly changes the organic matter content,
for example, simulation should be stopped and started again with better values
for those parameters.

SOLPOR, the soil porosity, is the fraction of the soil that can be filled
with water or air. The value for a soil can be calculated from the
bulk density, BD, the oven dry weight of a known volume of soil.
Assuming the solid density is 2.65 g/cnr:

SOLPOR = l-(BD/2.65) [11-21]

Values of porosity in the range of 0.3 to 0.5 are often available in
reports by the SCS. This value is used as POR in the NUT208 program
and P0R0S in the hydrology models.

FC, field capacity, is the fraction of the soil volume filled with water
after a day's drainage or in equilibrium with tensions of 0.1 to 0.3
bar. If measurements are impossible, some values in the range of
0.2 to 0.4 may be found in SCS reports. The value used here must be
compatible with the variable FUL used in the hydrology models.

0M, organic matter, is the percentage of the soil that is composed of bio-
logical residues. 0M is 1.724 times the percent total organic car-
bon in the soil. Values in the range of 0.1 to 2 for 0M or values
for total organic carbon are often given in reports from SCS. The
value for 0M must not be the same as that used for PER0G in the ero-
sion model, since 0M is the average in the root zone. If good in-
formation is unavailable, 0M can be set, realistically, as one half
of PER0G.

Initial Parameters

The user can select the method of nitrogen uptake calculations as given in
volume I, chapter 4. Plant nutrients in the surface soil layer and root zone
at the beginning of simulation can be measured, or estimated if measurements
are unavailable. These nutrient contents change with fertilizer, waste, and
residue applications as well as from plant uptake, leaching, denitrification,
and washoff. The model provides an accounting during the simulation, and only
initial values are needed.

OPT is 1 for nitrogen uptake to be simulated by plant growth and nitrogen
content. OPT is 2 when the normal probability curve is used to des-
cribe the nitrogen uptake.

295

SOLN and SOLP are the kilograms/hectare of soluble N and P in the surface
centimeter of soil. The initial value for these parameters is best
estimated by determining the equilibrium nitrate and phosphate con-
centrations in samples of the soil (CREAMS, vol. Ill, ch. 15). The
next best estimate is obtained using measured data for several
storms by fitting the relation PPMN=EXKN * CN where PPMN is the
parts per million concentration in the runoff, EXKN is the unitless
extraction coefficient, and CN is the parts per million concentra-
tion in the pore water of the surface centimeter of soil. CN is re-
lated to SOLN by

SOLN = 1/10 CN * POR [11-22]

where POR is the porosity.

Similar relations exist for phosphate. A default value in the
range of 0.01 to 0.4 for SOLN, SOLP, EXKN, and EXKP can be obtained
from information in CREAMS, volume III, chapter 14 and chapter 15.
Accuracy of the value for SOLP is most important, because the model
assumes that SOLP never drops below this value. The presence of
residues on the soil surface at the beginning of the simulation is
accounted for by having a nutrient addition on day 0.

N03 is the kilograms of Nitrate/hectare in the root zone. The initial
value should come from laboratory analysis of soil samples taken
from the root zone. A default value of 20 kg/ha can be used with
only a small effect for a long simulation period because this vari-
able is dynamic.

S0ILN and S0ILP are the contents of total nitrogen and total phosphorus in
the surface soil, kilograms of nutrient per kilogram of soil. These
values are available or can be estimated from SCS reports and soil
test results at State experiment stations. They range from 0.0005
to 0.003 for N and 0.0001 to 0.0013 for phosphorus.

EXKN and EXKP are unitless extraction coefficients whose estimation is
discussed in the preceding paragraph in connection with estimating
values for SOLN and SOLP.

AN and AP are enrichment coefficients, and

BN and BP are enrichment exponents for calculating the degree of N and P
enrichment in the sediment. These must be calculated from measured
values of N and P in sediments. Default values are 7.4 for the co-
efficients and -0.2 for the exponents.

RCN is the nitrogen concentration in rainfall in parts per million. The
concentration varies from slightly less to slightly more than 1 ppm.
A map in the description of the nutrient model shows how the nitro-
gen input in rainfall varies across the country (CREAMS, vol. I, ch.
4, fig. 1-16).

296

General Updateable Parameters

The nutrient model is structured such that dates of applicability are
specified by the user. Such date specification results in the program reading
at the appropriate time updateable information such as fertilizer additions.

PDATE is the first date (year and Julian day) on which the updateable pa-
rameters are valid.

CDATE is the last date (year and Julian day) on which the updateable pa-
rameters are valid. CDATE would be on a day prior to fertilizer ap-
plication as an example.

Updateable Parameters

Updateable parameters permit specification of the information that changes
with crop or for year-to-year changes for the same crop. Some parameters are
applicable to both options for nitrogen uptake. Parameters RZMAX, YP, DMY, and
POTM are required by both options.

NF is the number of nutrient additions (fertilizer, wastes, residues, and
so forth) that are made during the year.

DEMERG is the Julian date of plant emergence when nitrogen uptake starts.

DHRVST is the Julian date of harvest when nitrogen uptake stops.

Nitrogen uptake option 1— Nitrogen uptake by plants is calculated in this op-
tion by using the ratio of actual plant evaporation to potential plant evapora-
tion, AWU/PWU, and cubic coefficients to estimate the nitrogen content in the
crop dry matter.

RZMAX is the maximum depth of the potential root zone in millimeters.
This value is best obtained from field observations because many
fields have layers or conditions that limit root growth below normal
values for crops given in table 1-13 (CREAMS, vol. I, ch. 4) of the
model documentation. The value used here must be compatible with
depths used in the hydrology models.

YP is the kilograms/hectare potential yield of grain (seed cotton in the
case of cotton) for the crop grown under ideal conditions. Values
can be obtained from table 1-11.

DMY is the ratio of total dry matter yield (grain + stover + roots) to the
dry matter yield of grain.

POTM is the kilograms/hectare of potentially mineral izable nitrogen in the
root zone, which should be measured with laboratory tests. Default
values can be estimated from carbon or organic matter contents, us-
ing tables in CREAMS, volume III, chapter 13. Care must be taken
because values of carbon or organic matter in SCS reports are for
well managed soils and may be considerably higher than those for

297

poorly managed soils. Over estimation of this parameter can cause
over estimation of nitrate leaching. A low value is 50 kg/ha. POTM
is included in the updateable parameters to allow resetting to ac-
count for residue added after harvest.

AWU is the millimeters of actual water used by the crop and is the actual
transpiration accumulated for the year. Values of this parameter
are obtained from the output of the hydrology model.

PWU is the millimeters of potential water use by the crop and is the total
potential transpiration for the year. Preliminary runs of the hy-
drology model provides estimates of this parameter.

CI, C2, C3, and C4 are coefficients relating the nitrogen content of the
crop to its stage of growth as reflected in its amount of dry mat-
ter. These coefficients for corn, sorghum, wheat, cotton, and soy-
beans are given in table 3 of Smith and others (vol. Ill, ch. 13).

Nitrogen uptake option 2— The previously described updateable parameters RZMAX,
YP, DMY, and POTM are used with the option 2 method of estimating nitrogen up-
take by the crop. Nitrogen uptake calculations in this option are based upon
the number of days to reach 50% uptake, DOM, and the number of days between 50%
and 84% uptake, SD, determined from the normal distribution curve.

DOM is the number of days after emergence that half the nitrogen is taken
up and is equivalent to the mean of the probability distribution.

SD is the number of days required after 50% uptake to reach 84% uptake and
is equivalent to the standard deviation of the probability distribu-
tion. Estimates of DOM and SD for four crops are given in CREAMS,
volume III, chapter 13, table 5.

PU is the potential uptake of nitrogen, in kilograms/hectare, by the crop
under ideal conditions. Values are determined best from field stud-
ies, but estimates can be made as they are for YP.

The user can specify dates and rates of fertilization and depths of incor-
poration. The previously described parameter, NF, number of fertilizer appli-
cations, permit the user to make split applications. Fertilizer may be incor-
porated at planting time and a top-dress application of nitrogen may be added
later, for example.

DF is the Julian date that nutrients are applied to the field. If resi-
dues are on the field at the start of the simulation, a date of 0
can be used.

FN and FP are the kilograms/hectare of nutrients applied to the field on
each of the dates, DF. The content of nutrients in residues and
manures is given in tables I -11 and 1-12 of the nutrient model.

FA is the application factor that is the reciprocal of the depth of appli-
cation. Surface application is given a value of 1, while an appli-
cation that is mixed into the top 10 cm is given a value of 1/10.

Figure 11-35 schematically represents a data deck arrangement. The plant
nutrient and pesticide models are both included in the same computer program.

298

299

Table 11-38 is a list of parameters and definitions used in the nutrient
model, and it also gives the source and relative quality of estimates.

OUTPUT

Optional output is available to the user and is specified as FLGOUT, input
card 4. If only annual summaries are desired, FLGOUT =0. A sample of an an-
nual summary for plant nutrients is shown in figure 11-36. This summary shows
the total number of storms and total rainfall for the year, as well as the num-
ber of runoff-producing storms and total runoff. Unit nutrient losses are
shown for the elements and the values are accumulated for the year. Total nu-
trient losses are not added. For example, total phosphorus would be the sum of
phosphorus in runoff and phosphorus with sediment. Total nitrogen loss would
include nitrogen in runoff and nitrogen with sediment. Other elements in fig-
ure 11-36 include nitrogen uptake and mineralization, nitrate remaining in the
soil, rainfall nitrogen, nitrate leached, and denitrification. Maximum and
minimum values are given for nitrate leaching. Since leaching is difficult to
estimate, the extremes are given and actual leaching would be somewhere between
these values.

For storm values of nutrient losses, FLGOUT is set to 1. Figure 11-37
shows a sample of nutrient losses for a storm. Summarized input data from the
erosion pass file are shown at the top of this figure. The output data include
the type of data used for the annual summary, as well as soluble N and P avail-
able in the surface layer. Figure 11-38 shows output for a storm that did not
cause runoff. The data are abbreviated since runoff and erosion did not occur.
It is possible to have percolation from a storm that did not produce runoff,
and therefore, nitrate leaching is included. Storm output will help the user
to consider nutrient losses that might occur from storms shortly after applying
fertilizer. Concentrations of nitrogen and phosphorus in runoff (fig. 11-37)
are averages for the storm. Storm losses also are useful in considering sea-
sonal losses. They would be helpful in analyzing fertilizer-use efficiency as
well as nonpoint source pollution.

300

Table 11-39. — Inputs and parameters for pesticide submodel

Parameter

Definition

Source of estimate

Quality of estimate-'

' 3/

ARATE4'

Pesticide application
rate.

Recommendations on
label , farm re-
cords, table 11-40.

ID,
DEPINC.

- Depth of pesticide
incorporation.

Application recom-
mendation, experi
ence.

EF,

EFFINC.

- Efficiency factor for
incorporation.

Measurement, exper-
ience.

FF,

FOLFRC.

- Fraction on foliage

Model manual , ex-
perience, obser-
vations.

SF,

SOLFRC.

- Fraction on soil

Model manual , ex-
perience, obser-
vations.

FOLRES - •

• -Initial foliar residue

Experience, mea-
surement.

SOLRES - -

- Initial soil residue

Measurement, infer-

WSHFRC -

THRWSH -
H20S0L -

Cl/2.

HAFLIF.

DECAY.

EXTRCT.

KD.

Fraction of foliar
pesticide washed off.

Rainfall threshold for
washoff .

Pesticide solubility
in water.

Foliar pesticide half-
life.

Dissipation rate from
soil surface.

Extraction ratio, ratio
of soil : water in mix-
ing zone.

Distribution coeffici-
ent.

red from past man-
agement and pesti-
cide persistence.

Model manual , liter-
ature.

Judgment based on
canopy.

Handbooks, table II-
40, and table II-
41.

Model manual , liter-
ature, measurement.

Model manual

Model manual , liter-
ature, measurement.

Good, but may vary de-
pending on application
equipment and operator
care.

Good, but may vary de-
pending on soil condi-
tions.

Fair to good, depending
on soil conditions.

Fair to good, depending
on source of estimate.

Unknown, depends on
source of estimate.

Good if measured; poor
if inferred.

Good for limited number
of pesticides; fair to
unknown for others.

Probably fair, subjec-
tive.

Good to excellent for
most pesticides.

Fair to Good for limited
pesticides, but is
site- and condition-
specific.

Fair to good, but site-
and condition-specif-
ic, estimates from
bulk soil . Measure-
ments often under-
estimate.

Fair based on model per-
formance, but subjec-
tive.

Fair to good, but labor-
atory value may poorly
describe field behav-
ior.

1/ Excellent - known to be within few percent; Good - errors of 50% possible; Fair
error by factor of 2 possible; Poor - error by factor in excess of 2 possible.
2/ Notation used in documentation.
3/ Notation used in computer program.

301

ANNUAL SUMMARY FOR 1374

G7 STORMS PRODUCED 102. 2G CM. OF RAINFALL
14 STORMS PRODUCED 8.31 CM. OF RUNOFF

THE PLANT NUTRIENT LOSSES

NITROGEN IN RUNOFF 0.3214 KG/HA

PHOSPHORUS IN RUNOFF 0.3117 KG/HA

NITROGEN UITH SEDIMENT 17.G607 KG/HA

PHOSPHORUS WITH SEDIMENT G.7S51 KG-'KA

ACCUMULATED DRAINAGE 151. 1G MM

MINERALIZED N 14.2020 KG^'HA

N UPTAKE 144.3028 KG/HA

SOIL NITRATE 2.42G8 KG'HA

RAINFALL NITRATE 8.1808 KG/HA

ESTIMATE 1 NITRATE LEACHED 12.344G KG/HA

BURNS ESTIMATE 19,2052 KG/HA

ACCUMULATED DENITRIFICATION 23.877S KG -HA

Figure 11-36. — Sample output of annual sum-
mary from the plant nutrient component.

STORM INPUTS

DATE 74176

JULIAN DATE

RAINFALL 10.82

RUNOFF VOLUME 3.07

CM .

SOIL LOSS 4A70.50

KG/HA

ENRICH. RATIO 1 . SS

PERCOLATION 5 55

AVG . TEMP. 25.7 1

DEGREES C.

AVG. SOIL WATER 35

VOL/VOL

ACCUMULATED ET 53 A3

CM.

POTENTIAL ET 77 . 12

CM.

THE QUANTITY OF PLANT NUTRIENTS IN RUNOFF AND LEACHED

VALUES FOR STORM 74178

NITROGEN IN RUNOFF 0 KG/HA

NITROGEN IN RUNOFF 0 PPM

PHOSPHORUS IN RUNOFF . 1074 KG/HA

PHOSPHORUS IN RUNOFF 3500 PPM

NITROGEN WITH SEDIMENT 7.4875 KG/HA

PHOSPHORUS WITH SEDIMENT 2 8A20 KG/HA

DRAINAGE THIS STORM 35 46 MM

ACCUMULATED DRAINAGE 156.6°, MM

MINERALIZED N . A600 KG/HA

N UPTAKE 0 KG/HA

NITRATE LEACHED THIS STORM A 2111 KG/HA

SOIL NITRATE 124 5780 KG/HA

SOLUBLE N 0 KG/HA

SOLUBLE P 0A26 KG/HA

DENITRIFICATION 12. 4201 KG/HA

Figure 11-37. — Sample output of nutrient data for a

runoff-producing storm.

302

STORM INPUTS

DATE 74

171

JULIAN DATE

RAINFALL 1

CM.

RUNOFF VOLUME

CM .

SOIL LOSS

KG/HA

ENRICH. RATIO 2

PERCOLATION

CM.

AVG. TEMP. 25

DEGREES C.

AVG SOIL WATER

VOL/VOL

ACCUMULATED ET 51

CM.

POTENTIAL ET 72

*** NO RUNOFF -

LOSSES

***

DRAINAGE THIS STORM

0 MM

ACCUMULATED DRAINAGE

121

25 MM

MINERALIZED N

5656 KG/HA

N UPTAKE

0 KG/HA

NITRATE LEACHED THIS STORM

0 KG/HA

SOIL NITRATE

2A43 KG/HA

SOLUBLE N

8BA4 KG/HA

SOLUBLE P

2000 KG/HA

DENITRIFICATION

0 KG/HA

Figure 11-38. Sample output for plant nu-
trient data when runoff and nutrient
losses did not occur.

303

Chapter 4. THE PESTICIDE SUBMODEL

R. A. Leonard and J. D. NowlirV^

This submodel provides procedures to assess the effects of management op-
tions on potential pesticide losses in runoff. Its applicability is in making
relative comparisons among options. It is not designed to provide predictions
of pesticide concentrations in runoff to be used as an absolute value in making
water quality assessments. The model is for field-scale application and will
provide estimates of pesticide mass and storm-mean concentrations at the edge
of the field. The percentage of this quantity actually reaching and impacting
a body of water or stream is not addressed, and will depend on such factors as
location of the field with respect to receiving waters and properties of the
particular pesticide.

The impact of pesticides is caused largely by their concentrations in wa-
ter rather than their total mass. Pesticide concentrations are determined by
their rate of loss with respect to rate of runoff water and sediment and volume
of downstream receiving water. The submodel developed here does not describe
rate of pesticide transport within a single storm event. Experiments on small
plots have shown that pesticide concentration in runoff may decrease -several
fold from the beginning of runoff to the end, depending on storm duration and
mode of pesticide transport. Other experiments on field-sized or small, com-
plex-slope watersheds have shown that distinct wi thin-storm concentration pat-
terns are unusual except for pesticides that are transported by sediment. Con-
centrations within storms usually range between certain limits in an apparently
random way by factors of 2 or more, even up to 10. An explanation for this be-
havior is that the runoff material reaching the field edge originates at dif-
ferent locations within the field and has different times of travel to the mea-
suring point. In these situations, even accurate measurement of total storm
losses is difficult and representation of within-storm concentrations by models
is impossible without tremendous detail. In this submodel, use of daily or
storm totals generated by the hydrology and erosion submodels precludes any
within-storm description.

Details of model development are in CREAMS, volume I, chapter 5. Other
supporting documentation is provided in volume III, chapters 17, 18, and 19. A
potential user should consult this material for general familiarization with
the model .

1/ Soil scientist, Southeast Watershed Research Program, Athens, Ga., and-
computer programer, Agricultural Engineering Department, Purdue University,
West Lafayette, Ind., respectively.

304

MODEL STRUCTURE

The model is structured to account for multiple applications of the same
pesticide applied to soil or foliage. Different rates of dissipation or decay
can be used, if necessary, for that part of the chemical residing on foliage as
compared to that in soil. Movement of pesticide below the soil surface and out
of the runoff-active zone as a result of infiltrating water also is estimated
for potentially mobile compounds. Concentrations of pesticides in solution and
in sediment are computed, as well as total mass transported by each vehicle.
The initial pesticide residue concentration in the soil or on foliage, if any,
at the beginning of the modeling period is initially specified. Pesticide res-
idue remaining at the soil surface after each storm is computed, and the resi-
due and storm pesticide runoff are printed in the output.

MAJOR ASSUMPTIONS AND SIMPLIFICATIONS

In developing the model, many simplifying assumptions were required to re-
duce the description of complex systems and processes to a concept that could
be represented by simple mathematical expressions. The model user must be
aware of these assumptions and inherent limitations to avoid misapplication or
overinterpretation of the significance of the model outputs. Many assumptions
and limitations imposed and summarized here are discussed at length elsewhere
(CREAMS, vols. I and III).

Source of Pesticide in Runoff

The immediate soil surface is most active in supplying pesticide to run-
off. In interrill areas, extraction of pesticide may occur from a soil zone
only a few millimeters deep. Pesticide may be extracted, however, by active
rill erosion from a zone several centimeters deep. Extraction also may occur
as runoff water seeps through surface irregularities and furrows.

This model assumes an effective pesticide source zone of 1 cm deep at the
soil surface. Runoff concentrations are assumed to be proportional to pesti-
cide concentrations in this soil layer, expressed in units of micrograms per
gram (ppm). Initial surface concentrations after application are computed from
the rate of application and depth of incorporation. To specify a surface con-
centration, pesticides applied as a surface spray are assumed to be mixed uni-
formly with the 0- to 1-cm soil depth increment. Incorporated pesticide is as-
sumed to be uniformly or nonuniformly mixed throughout the depth of incorpora-
tion.

Rate of Pesticide Dissipation from Soil Surface

Pesticide is assumed to dissipate from, or decay in, the surface 0- to 1-
cm zone at a rate proportional to the amount present, as described by a simple
exponential function commonly known as a first-order rate expression. A single
parameter, referred to as the "decay constant," ks , is used in the function to
compute surface concentration as a function of time. This is a "lumped" param-
eter for degradation, volatilization, and other processes contributing to

305

pesticide dissipation from the soil surface. Assumptions and limitations in-
volved are discussed in detail in volume III. This simplification tends to un-
derestimate dissipation rates immediately after application and overestimate
dissipation rates after several weeks of pesticide contact with the soil. The
model provides no direct or prescribed way of incorporating the time variable
decay rate. However, the decay constant may be entered several times through-
out a model application period as an updateable parameter.

Rate of Pesticide Dissipation from Foliage

The dissipation of pesticide from foliage also is assumed describable by a
simple exponential decay function. In the model, foliar dissipation is des-
cribed by the parameter "half-life in days" or more correctly called half-con-
centration time, which is related to a decay constant, kf by: Half-life =
= 0.693/kf. T

Mechanism of Foliar Washoff

How washoff of pesticide from foliage contributes to observed runoff is
not well understood. For some pesticides, that part residing on foliage has
been described experimentally in terms of a fraction that can be dislodged and
a fraction that cannot be dislodged. Rainfall can remove part of the pesticide
described as "dislodgeable," depending on the pesticide and, probably, leaf
characteristics and time after application. In the model, a fraction of that
remaining as computed from the decay function is specified as "washoff frac-
tion." This part of the remaining pesticide is assumed to be moved from the
foliage to the soil surface when rainfall exceeds a "washoff threshold," which
is approximated by the amount of rainfall in centimeters that the plant canopy
can intercept and store as droplets on the surface of the leaf.

Vertical Transport of Pesticide from the Soil Surface

Pesticides that are mobile in soil, that is, soluble and not strongly ad-
sorbed, can be leached from the soil surface by infiltrating rainfall. Before
runoff concentrations are estimated for a storm, surface concentrations of
pesticide are reduced, depending on the amount of rainfall in excess of runoff
and initial wetting as a measure of flux through the surface layer of the soil.
For pesticides with solubilities greater than 1 ppm, a pesticide distribution
coefficient, Kj, is assumed to describe the availability of the pesticide for
transport by the infiltrating water. This procedure is approximate compared
with other more exact procedures that require detailed wi thin-storm informa-
tion. This simple method, therefore, may either overestimate or underestimate
vertical transport within a storm, depending on rainfall intensity and begin-
ning of runoff relative to rainfall.

Distribution of Pesticide Between Solution and Soil Phases

As indicated, a coefficient, K<-|, is assumed to describe the distribution
of pesticide between the water or solution phase and the adsorbed phase. This

306

coefficient is defined as the ratio of the concentration in soil (yg/g) to the
concentration in solution at equilibrium (yg/ml ) . Values of Kj normally are
assigned from equilibrium experiments in the laboratory using soil suspensions
containing added pesticide. In the model, the most serious assumptions regard-
ing the use of Kj are:

(1) Kd is independent of pesticide concentration. This assumption is
discussed in detail in volume III, chapter 19. Where this assumption is vio-
lated, the affinity of the soil or sediment for pesticide generally would be
underestimated at very low concentrations.

(2) Adsorption-desorption processes in soil are reversible, and equili-
brium is achieved rapidly. Many runoff experiments have shown that, with time
of contact in soil, pesticide becomes more difficult to displace in water; that
is, the apparent Kj increases. This observation may be related to both irre-
versible adsorption and to the dependence of Kj concentration. Equilibration
in the dynamic runoff stream probably is never achieved. If the desorption
rate is slow in relation to changes in the ratio of water to soil in the run-
off-active zone, solution extraction and transport will be overestimated.

Serious errors in applying and interpreting the model can be avoided if
Kd values are used to distinguish behavioral differences between major pesti-
cide classes (weakly adsorbed, moderately adsorbed, and so forth) as reflected
by Kd differences of an order of magnitude. When the model is used for rela-
tive comparisons, and when used in this manner, smaller differences in K^ may
be significant.

Pesticide Extraction Into Runoff

In developing the function relating concentrations of pesticide in runoff
water to concentration in the soil, assuming a value was necessary for the ra-
tio of soil:water in the mixing zone at the surface. Otherwise, it must be as-
sumed that the runoff water equilibrates with the pore water or extracts pesti-
cide from a mass of soil represented by the sediment yield. The total mass of
pesticide at the soil surface can be computed as a potential runoff source from
concentration and assumed depth. During a runoff event, however, all of this
pesticide does not react, or is not mixed, with runoff water. The extraction
ratio parameter, B, as defined, represents the effective soil:water ratio in
the mixed zone during a runoff event. The value of this parameter cannot be
measured directly and should be related to storm intensity, slope, and other
factors. A limited range of values for B is required, however, for satisfac-
tory predictions. Insufficient data are available to relate B or another
representation of the mixing zone to site and storm characteristics.

The model assumes that as the soil is mixed with runoff water at the soil:
water ratio specified, a distribution of pesticide between the solution and
soil phase is approached as approximated by Kj. Approximate equilibrium condi-
tions must be assumed. The solution concentration predicted at the field edge
is assumed to be the same as determined above. In the mixing zone, however,
the absorbed phase concentration computed is for the soil, not sediment. The
concentration in the sediment delivered at the field edge is assumed to be in-
creased by an enrichment factor or ratio reflecting preferential removal and

307

transport of clay and organic matter. An enrichment ratio is computed in the
erosion model based on particle characteristics of the sediment compared to
characteristics of the original soil.

The model has no mechanism for limiting the maximum mass of pesticide in
runoff during a single event. The surface concentration is reduced by vertical
transport before runoff. Since the surface concentration is reduced by the
amount in runoff only at the end of a runoff event, however, total runoff mass
may be overestimated in unusually large storms, that is, > 5-8 cm of runoff.

MODEL INPUTS AND PARAMETERS

Hydrologic inputs required are rainfall and runoff volume. These are ob-
tained from the hydrology model or input as observed data. Sediment yield is
obtained from the erosion model, experimental observations, or other estimates.
A hydrology pass file is used to generate an erosion pass file, which also con-
tains the hydrology data needed in the chemistry model. Figure 11-39 schemati-
cally represents the card deck from the erosion pass file. The figure was
given in the previous chapter for plant nutrients, but is repeated here for
user reference. The erosion model estimates enrichment factor for pesticide
transported by sediment. Table 11-39 identifies additional pesticide model
parameters and inputs required with suggested sources of estimate and expected
quality of the estimate.

(BLANK CARD FLAGS THE END OF THE FILE)

REMAINDER OF THE
STORM/HYDROLOGY/EROSION DATA
(I CARD/ EVENT)

fo rm at ( i6,F6. 2, F6. 2^6.2^6.2,12^6.2^6.2^6.4^6.3^6.3^6.3^6.3)

SDATE RNFALL RUNOFF 80 LOSS ENRICH DP PERCOL AVOTMP AVGSWC ACCPEV POTPEV ACCSEV POTSEV

Figure 11-39. — Schematic representation of a sample card deck arrangement and
format for the erosion/sediment yield pass file.

308

Table 11-38
Parameter Definition

SOILN Soil nitrogen

SOILP Soil phosphorus

EXKN Extraction coefficient

for nitrogen.

EXKP Extraction coefficient

for phosphorus.

AN Enrichment coefficient

for nitrogen.

BN Enrichment exponent

for nitrogen.

/\p Enrichment coefficient

for phosphorus.

BP Enrichment exponent

for phosphorus.

FC Field capacity

POR- Porosity

POTM Potential mineralization

for nitrogen.

RCN Concentration of nitrogen

in rainfall .

RZMAX Maximum depth

of root zone.

DOM Date of miduptake

SD Standard deviation

of uptake.

PU Potential nitrogen

uptake.

YP- Yield potential

Cp C~; C3; Plant nitrogen

C4 uptake coefficients.

— Nutrient model parameters

Source of estimate

Quality of estimate

Soil survey data;
lab analysis;
literature.

Analysis of runoff
data;
literature.

Analysis of erosion
data;
literature.

Soil survey data;
lab studies.

Lab analysis;
literature.

Field study;
soil survey.

Local information;
general information.

+ 40% Good for sampled

+ 20% soil series.
+ 100%

+_ 40% Dependent upon sam-

+ 20% pling scheme for

+ 100% unsurveyed soils.

100%

Do,

+ 300%

+ 100%

Do.

+ 300%

+ 30%

Do.

+ 300%

+ 30%

Do.

+ 300%

+ 30%

Do.

+ 300%

+ 30%

Do.

+ 300%

+ 30%

+ 15%

Excellent for point
samples; fair to
poor for varia-
bility in space.

+ 30%
+ 15%

Do.

+ 20%
+ 100%

Do.

+ 10%
+ 100%

Do.

20% . Good for cultivated
100% crops; poor for
weeds, range-
lands.

15% Generally not
30% available on a
local basis.

Local information;
general information.

+ 15%
+ 30%

Do.

Local information;
general information.

+ 15%

+ 30%

Do.

Local information;
general information.

+ 15%
+ 30%

Occasionally avail
able locally.

Manual

Good for crops mea
sured.

309

Selection of Input Values

The following discussion provides a guide to estimating input values and
parameters. Much of this information has been extracted from comprehensive re-
views and analyses in volume III of CREAMS. Where extensive data tables are
required, see volume III with suggestions, if appropriate, on how to use these
data. The discussions in volume III, help show how values were derived, possi-
ble errors, and how values may vary depending on site and condition. If avail-
able, use additional site-specific information from other sources rather than
average or generalized information in this publication.

Table 11-40 summarizes solubilities and application rates for some common-
ly used herbicides. Table 11-41 gives solubility for some common insecticides.
More complete tabulations as can be found in the handbooks referenced in these
tables.

310

Table 11-40.— Water

solubility (S0LH20) and application
ly used herbicides^'

rate (ARATE) of common-

Pesticide
trade
name

Pesticide

common

name

Water
solubility

Application rates

AMEX 820 A-820

Lasso- ALACHLOR

EVIK AMETRYNE

Amitrol-T AMITROLE

Dessicant ARSENIC ACID

AA trex — ATRAZINE

Balan BENEFIN

Basagran BENTAZON

Hyvar-X BROMACIL

Machete— BUTACHLOR

Sutan BUTYLATE

Bromex — CHLORBROMURON

Morex CHLOROXURON

2,4-D

DOWPON DALAPON

Banvel — DICAMBA

COBEX DINITRAMINE

DYMID, ENIDE DIPHENAMID

Karmex DIURON

Urab FENURON

Cotoran FLUOMETURON

Roundup GLYPHOSATE

PAARLAN ISOPROPALIN

Sencor METRIBUZIN

Daconate, Weed-Hoe MSMA

Telvar— - MONURON

Planavin NITRALIN

Ryzelan ORYZALIN

Ortho Paraquat PARAQUAT

Tordan PICLORAM

Tolban PROFLURALIN

Pramitol PROMETONE

Caparol PROMETRYNE

Ramrod PROPACHLOR

Milogard PROPAZINE

Pyramin PYRAZON

2,4,5-TP SILVEX

Princep — SIMAZINE

2,4,5-T 2,4,5-T

Randox TCBE

Treflan TRIFLURALIN

(ppm)

(lb/acre)

1.0

1 - 5

242

1 - 4

185

2-8

280,000

2-10

Freely

1.5

2 - 4

1.12 - 1.5

0.5 - 1.5

815

1.5-24

1.5 - 4

3 - 4

0.75 - 4

2.7

2 - 8

900

0.25 - 4

Very soluble

0.75 - 20

4,500

0.06 - 10

1/3 - 2/3

260

2-6

0.6-48

3,850

18 - 27

0.5 - 4

1.2%

1 - 4

0.11

1 - 2

1,220

0.25 - 1.0

Very soluble

2 - 3.8

230

4-48

0.6 •

0.5 - 1.5

2.4

0.75 - 1.75

Completely

0.25 - 1

430

1 - 8

0.1

0.5 - 1.5

750

10 - 60

0,48 - 2.75

580

3 - 6

8.6

1 - 4

400

2-4

140

0.75 - 16

2 - 4

238

0.5 - 16

2.6

0.5 - 2

1/ Hilton, H. L., R. W. Bovey, H. M. Hull, W. R. Mullison, and R. E.
Talbert. 1974. Herbicide Handbook of the Weed Science Society of America,
Third edition. Champaign, 111. 430 pp.

2/ Range for active ingredient.

311

Table 11-41. — Water solubility (S0LH20) of commonly used insecticides-^ -f

Insecticide Insecticide

trade common Water

name name solubility

Orthene ACEPHATE 65%

Guthion- AZINPHOSMETHYL 29

Bux BUFENCARB low

Sevin CARBARYL 40

Furadan CARBOFURAN 700

Lorsban — CHLORPYPIFOS 2

Spectracide, Diazinon- DIAZINON 0.004%

Di-Syston --DISULF0T0N 25

Dasanit FENSULFOTHION 1 ,600

Cythion MALATHION 145

Supracide METHIDATHION 240

Lannate, Nudrin METHOMYL 58,000

Metacide METHYL PARATHION 50 - 60

Methyl Parathion

Niran, Bladan— PARATHION 24

Thimet, Phorate-lOG— - PHORATE 50

Toxaphene TOXAPHENE 3

1/ Berg, G. L., (ed) . 1979. Farm Chemicals Handbook, Section D - Pesti-
cide Dictionary, Merster Pub. Co., Willoughby, Ohio. 316 pp.

2/ Lawless, E. W., T. L. Ferguson, and A. F. Meiners. 1975. Guidelines

for the disposal of small quantities of unused pesticides. U.S. Environmental

Protection Agency Technology Series. EPA-670/2-75-057. 331 pp.

Table 11-42 describes the pesticide input parameters and format, and a
schematic representation of an input data deck is shown in Figure 11-40.

312

Table 11-42. — Chemistry model input parameter file

Initial General Parameter Inputs

Card 1-3.

Card 4.

Card 5.

TITLE ( )
TITLE

Three lines of 80 Characters each for alphanumeric
information to be printed at the beginning of the out-
put, format (20A4)

BDATE, FLGOUT, FLGIN, FLGPST, FLGNUT

BDATE The beginning date for simulation. It must be less
than the first storm date (SDATE) . (Julian date), e.g.
73138

FLGOUT 0 for annual summary output

1 for annual and monthly summary output

2 for individual storm and all summary output

FLGIN 0 if the Storm Hydrology input is in English units and
will need to be converted to metric
1 if the values are already in metric units

FLGPST 0 if there will be no Pesticide inputs
1 if there will be Pesticide simulation

FLGNUT 0 if there will be no Plant Nutrient inputs
1 if there will be Plant Nutrient simulation

SOLPOR, PC, CM

SOLPOR Soil porosity (cc/cc) , e.g. 0.41

FC Field capacity (cc/cc), e.g. 0.32

CM Organic matter available for denitrif ication (% of soil
mass) , e.g. 0.65

Initial Pesticide Inputs

Card 6. NPEST, PBDATE, PEDATE
NPEST

PBDATE

Number of pesticides, e.g. 2, MAX of 10

If blank the pesticides portion of the model is

bypassed .

Date the model begins to consider pesticides (Julian
date) , e.g. 74120

313

Table 11-42. — Chemistry model input parameter file—continued

PEDATE Date the model stops considering pesticides (Julian
date) , e.g. 75365

Initial Plant Nutrient Inputs

Card 7. OPT

OPT 1 for option one nitrogen uptake

2 for option two nitrogen uptake

Card 8. SOLN, SOLP, N03, SOILN, SOILP, EXKN, EXKP, AN, BN, AP

SOLN Soluble nitrogen (kg/ha), e.g. 0.2

SOLP Soluble phosphorous (kg/ha), e.g. 0.2

N03 Nitrate (kg/ha), e.g. 20.0

SOILN Soil nitrogen (kg/kg), e.g. 0.00035

SOILP Soil phosphorous (kg/kg), e.g. 0.00018

EXKN Extraction coefficient for nitrogen, e.g. 0.0576

EXKP Extraction coefficient for phosphorous, e.g. 0.07

AN Enrichment coefficient for nitrogen, e.g. 16.8

BN Enrichment exponent for nitrogen, e.g. -0.16

AP Enrichment coefficient for phosphorous, e.g. 11.2

Card 9. BP, RCN

BP Enrichment exponent for phosphorous, e.g. -0.146

RCN Concentration of nitrogen in rainfall (mg/1) , e.g. 0.8

Updateable General Parameter Inputs

Card 10. PDATE, CDATE

314

Table 11-42. — Chemistry model input parameter file— continued

PDATE

CDATE

First date that the following chemical parameters are
valid (Julian) , e.g. 73138

The program doesn't read in a value for PDATE. PDATE
is only used as an aid in putting together the data
file.

Last date that the following chemical parameters are
valid, for example one day before the next pesticide
application or one day before a change in the plant
nutrients parameters (Julian) , e.g. 73131

NOTE: A card 10. should always be the first card in a
new set of updateable parameters.

Updateable Pesticide Inputs

Card

11.

APDATE
APDATE

Card

12.

PSTNAM ( )
PSTNAM

Card

13.

APRATE,
WSHTHR

APRATE

DEPINC

EFFINC

FOLFRC

SOLFRC

FOLRES

Date the pesticide is applied (Julian date) , e.g. 73121
if blank, cards 12-14 are not read

The pesticide name, up to 24 characters, format (6A4) ,
e.g. ATRAZINE

DEPINC, EFFINC, FOLFRC, SOLFRC, FOLRES, SOLRES, WSHFRC,

Rate of application (kg/ha), e.g. 3.36

Depth of incorporation (cm), e.g. 1.0

Efficiency of incorporation, e.g. 1.0

Fraction of pesticide applied to the foliage, e.g. 0.0

Fraction applied to the soil, e.g. 1.0

Amount of pesticide residue on the foliage prior to
this application (ug/g) , e.g. 0.0

SOLRES Amount on the soil prior to this application (ug/g) ,
e.g. 0.0

WSHFRC Eraction on the foliage available for rainfall washoff,
e.g. 0.0

WSHTHR Rainfall threshold for foliage washoff (cm), e.g. 0.0

315

Table 11-42. — Chemistry model input parameter file—continued

Card 14. SOLH20, HAFLIF, EXTRCT, DECAY, KD

SOLH20 Water solubility (PPM), e.g. 33.0

HAFLIF Foliar residue half life (days), e.g. 0.0

EXTRCT Extraction ratio, e.g. 0.1

DECAY Decay constant, e.g. 0.10

KD KD, e.g. 2.0

Cards 11-14 are repeated for each pesticide (NPEST, card 6). If the
application date (APDATE, card 11) is blank then cards 12-14 are omitted for
that pesticide and the old values, including APDATE, are retained. This is
use full when one of the pesticides is to be reapplied but others are not. If
more than one pesticide is applied and the pesticides are applied on different
dates, blank, cards must be inserted at the appropriate places in the file for
each pesticide not being applied with this update. The following example is
given for clarification.

Assume 3 pesticides (NPEST, card 6) are applied with the following application
dates:

Atrazine - 3/20/74 (74079) , 3/27/75 (75086)

2,4-D -4/15/74 (74105), 4/12/75 (75102)

Parathion - 6/13/74 (74164) , 7/05/74 (74186) ,

6/20/75 (75171), 7/21/75 (75202)

The following cards 10-14 would be used:

Card 10
Card 11
Card 12
Card 13
2 blank
Card 10
1 blank
Card 11
Card 12
Card 13

1 blank
Card 10

2 blank
Card 11
Card 12
Card 13
Card 10
2 blank
Card 11
Card 12

74000 74104

74079
Atrazine
and 14: appropriate data
card ll's for 2,4-D and Parathion

74105 74163
card 11 for Atrazine

74105
2,4-D
and 14: appropriate data
card 11 for Parathion

74164 74185
card ll's for Atrazine and 2,4-D

74164
Parathion
and 14: appropriate data

74186 75085
card ll's for Atrazine and 2,4-D

74186
Parathion

316

Table 11-42 Chemistry model input parameter fil e--continued

Card 13 and 14: appropriate data

Card 10: 75086 75101

Card 11: 75086

Card 12: Atrazine

Card 13 and 14: appropriate data

2 blank card ll's for 2,4-D and Parathion

etc

Updateable Plant Nutrient Inputs

Card 15. NF, DEMERG, DHRVST

NF number of fertilizer applications, e.g. 2

DEMERG Date of plant emergence (Julian date, no year), e.g.

141

DHRVST Date of plant harvesting (Julian date, no year) , e.g.
305

When no new Plant Nutrient values are to be read card 15 should be left
blank. The program will then skip reading the remaining Plant Nutrient param-
eters.

por option One Nitrogen Uptake

Card 16. RZMAX, YP, DMY, POTM, AWU, PWU

RZMAX Maximum depth of the root zone (mm), e.g. 450.0

YP Potential yield (kg/ha), e.g. 5700.0

DMY Dry matter yield ratio, e.g. 2.5

POTM Potential mineralizable nitrogen (kg/ha), e.g. 47.0

AWU Actual water use (mm), e.g. 570.0

PWU Potential water use (mm), e.g. 780.0

Card 17. CI, C2, C3, C4

C1,C3 Cubic coefficients, e.g. 0.0209, 0.0128

C2,C4 Cubic exponents, e.g. -0.157, -0.415

317

Table 11-42. — Chemistry model input parameter file—continued

For Option Two Nitrogen Uptake

Card 16. RZMAX, YP, DMY, POTM, DOM, SD, PU

RZMAX Maximum depth of the root zone (mm) , e.g. 450.0

Potential yield (kg/ha), e.g. 5700.0

Dry matter yield ratio, e.g. 2.5

Potential mineralizable nitrogen (kg/ha), e.g. 47.0

DMY

POTM

DOM

SD
PU

Date of mid point in nitrogen uptake cycle (days) , e.g.
73.0

Standard deviation of DOM (days), e.g. 30.0

Potential nitrogen uptake (kg/ha) , e.g. 250.0

Both Options Continue

Card 18. DF(1)

DF Date of fertilizer application (Julian date) , e.g.

73131

Card 19. FN(1) , FP(1), FA(1)

EN Nitrogen applied (kg/ha), e.g. 28.0

FP Phosphorous applied (kg/ha) , e.g. 28.0

FA Surface fraction of application, e.g. 0.1

Cards 17 and 18 are repeated for each application of fertilizer (NF, card
15). A maximum of 20 applications can be read in one update.

318

Table 11-42. — Chemistry model input parameter file—continued

A sample data file for the Control Parameters for pesticides follows,
will help demonstrate the file structure.

CARD
NO

6
10
11
12
13
14
11
10
11
11
12
13
14
10
11
12
13
14
11
10
11
11
12
13
14
10
11
12
13
14
11
10
11
11
12
13
14
10

CHEMISTRY PARAMETER DATA

73138

0.410

73121

ATRAZINE

3.3G0

33.0

0
73132

PARAQUAT
2.049

500000.0

74121

ATRAZINE

3.3G0

33.0

0
74132

PARAQUAT
2.043

500000.0

75121

ATRAZINE

3. 360

33.0
0

0
75132

PARAQUAT
2.049

500000.0

0
0.320
74120
73131

1.000
0.0

74120

PESTICIDES PARAMETERS - GEORGIA PIEDMONT
MANAGEMENT PRACTICE ONE
CONTINOUS CORN - COHUENTIONAL TILLAGE
0 1 0
0.G50
753G5

1.000 0.000
0.1000 0.1000

1.000
2.0

0.000 0.000 0.000 0.000

1.000 1.000 0.000 1.000 0.000 0.000 0.000 0.000

0.0 0.1000 0.0070100000.0
74131

1.000 1.000 0.000 1.000 0.000 0.000 0.000 0.000
0.0 0.1000 0.1000 2.0

75120

1.000 1.000 0.000 1.000 0.000 0.000 0.000 0.000

0.0 0.1000 0.0070100000.0
75131

1.000 1.000 O.C00 1.000 0.000 0.000 0.000 0.000
0.0 0.1000 0.1000 2.0

753GG

1.000 1.000 0.000 1.000 0.000 0.000 0.000 0.000
0.0 0.1000 0.0070100000.0
0

319

320

Appl i cation Rate

The desired pesticide rate for a given application usually is specified
within certain limits by the registration data on the label or is obtained as
recommendations from the supplier or extension specialists. The number of ap-
plications for some pesticides, particularly foliar-applied insecticides, will
depend on extent of insect infestation or established spray schedules. Appli-
cation rate is input in units of kilograms/hectare. See table 11-40 for ranges
of application of some common herbicides.

Depth of Pesticide Incorporation

Pesticides often are incorporated by double-disking, rotary tillers, and
other equipment for harrowing or smoothing the soil surface. Depth of pesti-
cide incorporation will depend on the type of tillage equipment used and soil
conditions. Depth of incorporation normally ranges from about 8 to 15 cm (3 to
6 in). When the pesticide is incorporated, select the depth based on the till-
age equipment used. For surface-applied chemicals, a value of 1 cm is input as
the incorporation depth since the surface is defined arbitrarily as having a
depth of 1 cm.

Efficiency Factor for Incorporation

Most incorporation devices do not mix the applied pesticide uniformly
throughout the entire depth. The concentration remaining at the surface may be
significantly higher than at lower depths. Injected pesticides may have a low
surface concentration due to their placement below the surface. The efficiency
factor can be used to adjust the surface concentration based on known patterns
of incorporation. If an incorporation device leaves a concentration in the
surface of twice that achieved by uniform mixing, for example, an efficiency
factor equal to half the incorporation depth could be used. For injected pest-
icides, an efficiency factor of less than one will reduce the surface concen-
tration in proportion. Since this type of information usually is unavailable,
a value of 1 would be input with the assumption that uniform mixing was achie-
ved.

Fraction on Soil and Foliage

When crops are treated with pesticides applied to the plant canopy, some of
the application, depending on degree of canopy closure, will reach the surface
of the soil directly, some will remain on the foliage, and the rest will be
lost by drift and volatilization. At full canopy, about 75 + 20% and 50 + 20%
of the ground and aerial applications, respectively, reach the canopy (CREAMS,
vol. Ill, ch. 18). If the amounts reaching soil directly are assumed negligi-
ble at full canopy, about 25 to 50% can be lost by drift and volatilization
during application. For incomplete canopy, the fraction reaching soil should
be somewhat proportional to the extent of ground cover although insufficient
information is available to provide any functional relationship. The actual
distribution between soil, foliage, and off-target loss will be highly variable
and dependent on atmospheric conditions, path of application, and canopy char-
acteristics. If site-specific information is unavailable, at full canopy clo-
sure use 0.4 to 0.6 on foliage for aerial applications and 0.7 to 0.8 on foli-

321

age for ground applications. Assume an insignificant fraction reaching the
soil. For less than full closure, use a fraction for soil interception in pro-
portion to exposed ground surface. For example, suppose an aerial application
is made to cotton that, on projection, covers 50% of the ground surface. The
fraction on foliage would be 0.3 and the fraction on soil would be 0.3, with
the rest, 0.4, assumed as off-target losses.

Initial Foliar Residues

Pesticides normally dissipate from foliage such that a residue will not be
present at the beginning of a model application period. This option, is pro-
vided, however, so that the model can be applied on any date. To estimate an
initial residue from a previous application, assume interception fraction, as
was suggested, and use equations given in volume III, chapter 18, to estimate
dissipation with time. Rates of foliar dissipation are discussed in a follow-
ing section. The value input should be in units of milligrams of pesticide per
square meter of ground surface. Initial residue can be determined best by di-
rect measurement, but this procedure usually is not practical except for re-
search.

Initial Soil Residue

As for foliar residue, the amount of pesticide present in soil at the be-
ginning of a model application period is best determined by sampling and analy-
sis. Little residue of nonpersistent pesticides would be expected at the be-
ginning of a growing season. When persistent pesticides, such as organochlor-
ines, have been used for several years on a site, however, a significant resi-
due will be present. If sampling and analysis cannot be accomplished, publish-
ed data should be sought on residues in the soils of the area. The input value
should be in units of micrograms per gram (ppm). If the initial residue cannot
be determined by measurement or cannot be estimated from published information
such as that found in the Pesticide Monitoring Journal, levels of initial resi-
due may be estimated by using the values in volume III, chapter 17, if past ap-
plication history is known. First-order decay may be assumed, or an equation
of best fit may be used, such as that in volume III, chapter 17 on observed
pesticide persistence.

The initial residue parameter also provides a device for updating the con-
centration of pesticide in the surface of the soil as a result of redistribu-
tion caused by major tillage. Persistent pesticide may accumulate at the soil
surface during an application season. This accumulated residue would be pre-
dicted as output from the model. At the time of tillage, a new value for the
concentration at the surface of the soil can be computed, based on the accumu-
lated residue and tillage depth, and can be entered as an initial soil residue
for a new model application period.

Foliar Washoff Threshold

This parameter estimates the amount of rainfall required to exceed the ca-
pacity of the canopy to intercept and retain rainfall as droplets on the leaf

322

surfaces. Once this amount of rainfall is exceeded, pesticide washoff is as-
sumed. The value of this parameter probably ranges from about 0.1 cm to 0.3 cm
for a dense crop canopy.

Washoff Fraction

Little information is available on extent and patterns of pesticide wash-
off from foliage. The efficiency of the washoff process may be related to sev-
eral factors. Information in volume III, chapter 18, suggests that rainfall
can remove about 60% of the dislodgeable residue of most pesticides. Organoch-
lorines, and possibly other pesticides, however, are exceptions. Less than 10%
of these compounds is removed by rainfall. Values of 0.6 to 0.7 are suggested,
therefore, for all except the organochlorines, where values in the range of
0.05 to 0.1 should be used.

Water Solubility

Pesticide solubilities can be found in many handbooks on pesticide proper-
ties. In the model, solubility serves two functions. If solubility is < 1
ppm, the vertical transport computation is bypassed. Secondly, the predicted
runoff concentration in solution is compared to solubility. If solubility is
exceeded, the solution concentration is limited to the water solubility. Solu-
bility is, therefore, a critical input parameter only for the relatively insol-
uble pesticides. Solubilities of some common pesticides are given in tables
11-40 and 11-41.

Foliar Residue Half-Life

Consult volume III, chapter 18, for half-life values of pesticides on fol-
iage. Pesticides generally are not as persistent on foliage as in soil.

Extraction Ratio

This parameter describes the efficiency of the runoff stream in removing
or extracting pesticide. Conceptually, it is the ratio of soiltwater in the
mixing zone. Tests with the model indicate that values in the range of 0.05 to
0.2 are needed — the higher values for conditions of excessive runoff and ero-
sion. Predicted runoff concentrations of those pesticides transported entirely
in solution vary in direct proportion to the value of the extraction ratio. As
sediment transport becomes more significant, sensitivity to this parameter de-
creases. A value of 0.1 gives adequate prediction in most situations.

Soil Decay Constant

Values of rate constants, ks , are tabulated in volume III, chapter 17 for
the assumed expoential decay function applied to several pesticides and condi-
tions. Because dissipation rates are affected by climatic factors, the results
of individual experiments also should be reviewed before making a final selec-

323

tion (vol. Ill, ch. 17). Note that decay constants for surface and subsurface
(bulk soil) are given in volume III, chapter 17, if data were available. Many
pesticides dissipate more rapidly at the surface of the soil than from the soil
bulk. The ks values for surface dissipation are more appropriate for runoff
prediction, but more results have been reported on persistence in the soil
bulk. Where ks values are given for soil bulk but not for surface, differences
reported for similar compounds may be used in making a subjective judgment on
how the surface ks might differ from the reported bulk soil ks.

Additional information is provided in volume III, chapter 17, on how ks
values can be estimated based on properties of the pesticides and their envir-
onment. In addition to a better perspective of factors influencing dissipation
rates, methods are provided by which ks values can be estimated where little
experimental data are available.

In some instances, the first-order decay equation poorly describes dissi-
pation of a pesticide. Another alternative is suggested in chapter 17, volume
III, whereby pesticide concentration as a function of time can be obtained from
equations fitted to experimental data. No direct method is provided in the
present model for substituting these equations for the first-order decay equa-
tion. The ks values can be updated, however, using different values for dif-
ferent times after application. A best-fit equation could be used to compute
ks values for shorter time segments of the linear log c vs. t relationship as-
sumed. Since all equations of best fit are not incorporated in this version of
the model, a user should consult the author of chapter 17, volume III when nec-
cessary.

Distribution Coefficient Kj

Chapter 19, volume III, discusses how K<j is determined, the factors af-
fecting its value for different pesticides and soils, and how to estimate ft
for a specific situation. Tables 1 through 4 list mean K<j with standard devia-
tions for several pesticides. These tables also provide for estimating Kj as a
function of soil texture and organic matter content, thus tying k^ to both the
pesticidal properties and controlling site-specific characteristics of the
soils. Additional relationships for estimating Kj are based on observed soil
thin-layer chromatography and pesticide solubility.

Some assumptions are discussed for using Kj to predict distribution of
pesticide between solution and adsorbed phases. Figure 1 (volume III, chapter
19), shows how the apparent K^ can vary with pesticide concentration if the ad-
sorption relationship or isotherm is nonlinear. Users should compare potential
errors due to linearity and other assumptions in relation to the accuracy of
required output to achieve the objectives of their simulation. Since the ef-
fect of these assumptions on the validity of model output is uncertain, Kj val-
ues for an order of magnitude might be warranted when distinguishing major be-
havioral differences. Expressing K<j values explicitly as per reference may be
useful to analyze certain problems or situations, using model simulations to
compare effects of different management alternatives on the same site.

324

OUTPUT

The user can specify the type of output from the pesticide model by the
input on card 4: FLGOUT = 0 for annual summary only, FLGOUT = 1 for monthly
and annual summary, or FLGOUT = 2 for storm output as well as monthly and
annual summaries. This enables users to select the best output for their
problems. If a potential toxicity problem exists, storm output would be
needed, whereas an overall assessment of the pesticide losses could be
determined from the annual summary.

Figure 11-41 shows an annual summary output for a situation where six
pesticides were applied and a seventh pesticide was applied in previous years.
A storm summary of rainfall and runoff for the year is shown at the top of the
figure which gives the total mass of pesticide in water and with sediment.
Loss of pesticide as a percentage of that applied is shown also. Only the to-
tal mass is shown for toxaphene, which was not applied during the year and the
percentage of application shows residue. Figure 11-42 shows sample output of
monthly summaries for the same seven pesticides.

Figure 11-43 shows output for a single storm event when there was no run-
off or pesticide loss. Figure 11-44 shows the model output for a storm event
that resulted in runoff, erosion, and pesticide loss. The pesticide numbers in
figure 11-44 correspond to order of input and the order for the annual summary
(fig. 11-41). Concentrations in water and sediment are averages for the storm.

RNNURL SUMMRRY FOR 1A74

107 STORMS

PRODUCED 17R.

CM. OF RRINFRLL

49 STORMS

PRODUCED 63.

CM OF RUNOFF

THE

PESTICIDE

LOSSES

PESTICIDE

TOTRL MRSS

PERCENT OF

NAME

G/HR

APPLICATION

FLUOMETURON

B.26

.55

TRIFLURRLIN

.05

.01

MSMR

26B 14

17 SB

DIURON

METHYL PARATHION

. 66

EPN

1AA.71

3 AA

TOXRPHENE

2A6 31

RESIDUE

Figure 11-41. — Sample output of annual summary of
pesticide component where six pesticides were
applied and the seventh pesticide carried over
from previous years.

325

MONTHLY SUMMARY FOR JUL. 1A74

7 STORMS PRODUCED
2 STORMS PRODUCED

14.85 CM. OF RAINFALL
1 85 CM . OF RUNOFF

THE PESTICIDE LOSSES

PESTICIDE

TOTAL

MASS

PERCENT OF

NAME

G/HA

APPLICATION

FLUOMETURON

.04

RESIDUE

TRIFLURALIN

.05

RESIDUE

MSMA

RESIDUE

DIURON

.27

METHYL PARATHION

. 15

. 01

EPN

1 .66

TOXAPHENE

RESIDUE

MONTHLY SUMMARY FOR AUG.

15 STORMS PRODUCED
5 STORMS PRODUCED

16.56 CM. OF RAINFALL
1 . 46 CM . OF RUNOFF

THE PESTICIDE LOSSES

PESTICIDE

TOTAL

MASS

PERCENT OF

NAME

G/HA

APPLICATION

FLUOMETURON

.00

RESIDUE

TRIFLURALIN

RESIDUE

MSMA

RESIDUE

DIURON

. 01

RESIDUE

METHYL PARATHION

.01

EPN

.45

5 . 77

TOXAPHENE

RESIDUE

Figure 11-42. — Sample output of monthly summaries
for the pesticide component where six pesti-
cides were applied and the seventh pesticide
carried over from previous years.

326

STORM INPUTS

DOTE 741A2

JULIAN DRTE

RAINFALL

CM.

RUNOFF VOLUME

CM.

SOIL LOSS

KG/HA

ENRICH. RATIO 2

.00

PERCOLATION

CM.

AVG. TEMP. 27

.57

DEGREES C .

AVG. SOIL WATER

VOL/VOL

ACCUMULATED ET 55

.50

POTENTIAL ET S0

. 77

CM.

*** NO RUNOFF -

LOSSES ***

Figure 11-43. — Sample output from the pesticide component
for a single storm event that did not produce runoff.

STORM INPUTS

DATE

741R0

JULIAN DATE

RAINFALL

5 . S7

CM.

RUNOFF VOLUME

1 31

CM .

SOIL LOSS

726 S5

KG/HA

ENRICH. RATIO

2 4S

PERCOLATION

CM.

AVG. TEMP.

27 14

DEGREES C

AVG. SOIL WATER

.28

VOL/VOL

ACCUMULATED ET

55 SI

CM .

POTENTIAL ET

BA . 2A

QUANTITY OF PESTICIDE IN RUNOFF
VALUES FOR STORM 741A0

PEST. CONC. AVA.
NO RESIDUE

UG/G

CONC IN
WATER
UG/ML

MASS IN CONC. IN MASS IN
WATER SEDIMENT SEDIMENT
G/HA UG/G G/HA

TOTAL REMAIN.
MASS RESIDUE
G/HA UG/G

0
00
00

0004

0026

0478
3575

0001

0012

0
0000
000A

047A

53B2

Figure 11-44. — Sample output from the pesticide component for a runoff -

producing storm.

327

MODEL APPLICATION

Selection of best management practices rarely will hinge around solving
only a single potential problem affecting the soil resource or water quality or
both. A balance will be sought among total production, production efficiency,
net profits, protection of the resource base, and need and potential for im-
proving downstream water quality. For most constituents in water draining from
agricultural fields, including pesticides, no absolute standards or criteria
have been set as goals or requirements that must be met. Model output cannot
be compared, therefore, for selecting management options that meet a fixed set
of criteria. If this were possible, predictions and criteria also must deal
with probabilities of occurrence and the permissible or reasonable level of en-
vironmental risks.

Whenever a toxin is used widely, some risk is incurred to at least part of
the environment. Acute toxicity problems are identified more easily by their
effects than long-term chronic exposure. Dangers of long-term exposure to very
low levels of chemicals in the environment are not well understood, nor is
there general agreement on the extent of danger. Therefore, the basic question
of how much pesticide runoff consititutes a problem and how much it should be
reduced cannot be answered at this time, and is beyond the scope of this dis-
cussion. The general philosophy in the environmental community is that reduc-
tion of all off-target losses of pesticides to some practical minimum is desir-
able. This is not to say, however, that some management option shows potential
for reducing pesticide runoff should necessarily be selected over another op-
tion. All factors must be considered, including reduction of soil and plant
nutrient losses, effects on production, costs, and net return; and potential
problems caused by the pesticide. In using nonpoint source pollution models,
therefore, the planner for land use and water quality must examine and rate
management options with uncertainties of the issue as well as uncertainties in
the model outputs.

An example of how the model could be used is to compare relative losses of
pesticides under different management schemes designed to limit sediment yield.
Table 11-43 shows results of a simulation for a hypothetical situation where it
was assumed that .3 cm rainfall occurred on days 5, 10, 15, 20, and 25 after
pesticide was applied on the surface at the rate of 3 kg/ha. Each storm was
assumed to produce 1 cm runoff and 500 kg/ha sediment yield. Pesticides with
Kd's of 5 and 5,000 were considered. Both pesticides were assumed to have de-
cay constants, ks , of 0.10, and a sediment enrichment factor of 2.0 was assumed
in all events. In terms of total mass, the pesticide with a Kj = 5 was trans-
ported almost totally in the water phase, whereas the pesticide with a fy of
5,000 was transported by sediment. At the assumed level of sediment produc-
tion, total losses of the sediment-transported pesticide were less than those
predicted for the water-transported pesticide. Pesticide losses would be simi-
lar in each situation if sediment production was increased by a factor of about
4. Total loss of the pesticide with a K,j of 5 would not be changed signifi-
cantly, however, by increased or decreased sediment yield unless the volume of
runoff also was changed. Sample model runs on actual situations are given in
chapter 5 of this volume and may be examined for additional illustrations of
model use.

328

Table 11-43. — Pesticide in runoff predicted for hypothetical situation of 3 cm
rainfall, 1 cm runoff, and 500 kg/ha sediment yield on days indiated^'

Days after Concentration

Concentration

Total mass

Percent in

appl ication in water

sediment

in runoff

water

(ppb)

(ppm)

(grams)

(%)

K . = 5
5 a 670

6.70

70.3

10 335

3.35

35.3

15 170

1.70

17.8

20 85

0.85

8.9

25 43

0.43

4.5

K. = 5,000
5 d 2.4

24.2

12.4

10 1.5

14.7

7.4

15 0.9

9.0

4.6

20 0.5

5.4

2.8

25 0.3

3.3

1.7

1/ R = 3.0 kg/ha, k = 0.10,

enrichment factor =

2, extraction

coefficient

= 0.17 s

329

Chapter 5. EXAMPLE APPLICATIONS FOR TYPICAL FIELD SITUATIONS

G. R. Foster, M. H. Frere, W. G. Knisel, R. A. Leonard, A. D. Nicks
J. D. Nowlin, R. E. Smith, and J. R. Williams-''

INTRODUCTION

This chapter cites three typical field situations to show how parameter
values are obtained for a real -world problem. Limited interpretive information
will help the user understand the significance of specific aspects of the
CREAMS model and its parameters. The three typical field situations represent
different physiographic areas of Georgia Piedmont, Mississippi Delta, and west-
ern Tennessee. These examples typify gently rolling topography; flat, land-
formed topography; and steep slopes with long, slender fields.

The management practices for sample computer runs may not be recommended
by the SCS or acceptable by farmers, but the procedures are valid and should
help the user understand the model operations. Two management practices are
considered for each location. Table 11-44 shows the three locations, manage-
ment practices (MP1 and MP2), and model components.

DESCRIPTION OF APPLICATION SITES

Georgia Piedmont

The topography of the Georgia Piedmont field (fig. 11-45) is typical of
Piedmont cropland. Drainage from the field is restricted at the fence line and
causes in some temporary ponding of runoff. The soil is Cecil sandy loam with
a depth of 24 in to the B2 horizon. Continuous corn is assumed for the crop.

Two management practices in table 11-44 are (1) MP1 , conventional tillage
with rows running across the drainage, more or less on the contour in the upper
end and (2) MP2, modified tillage with a grass waterway extending approximately
two-thirds of the field length. Conventional tillage consists of spring mold-
boarding, disking twice, planting, and cultivating twice. Modified tillage
consists of chiseling, disking, planting, and not cultivating. Plant nutrient
application consists of the locally customary application of 140 kg/ha of

1/ Hydraulic engineer, USDA-SEA-AR, Lafayette, Ind. ; soil scientist, USDA-
SEA-AR, New Orleans, La.; hydraulic engineer, USDA-SEA-AR, Tucson, Ariz.; soil
scientist, USDA-SEA-AR, Athens, Ga.; agricultural engineer, USDA-SEA-AR, Chick-
asha, Okla.; computer technician, Purdue University, West Lafayette, Ind.;
hydraulic engineer, USDA-SEA-AR, Fort Collins, Colo.; and hydraulic engineer,
USDA-SEA-AR, Temple, Tex.

330

Table 11-44.— Typical field situations for sample model runs

Location

Model Component Georgia
Component Option Piedmont

Hydrology Option 1
Option 2

Erosion

Nutrients

Pesticides

Method 1
Method 2

MP1

Mississippi
Delta

MP1

MP1
MP2

MP1

western
Tennessee

MP1
MP2

MP1

1/ MP1 is management practice 1; MP2 is management practice 2.

100

200

SCALE IN FEET

CONTOUR INTERVAL 1 FOOT
ELEVATION M.S.L.

DRAINAGE
OUTLET

Figure 11-45. — Topographic map for Georgia Piedmont field.

331

nitrogen and 28 kg/ha of phosphorus. At planting time, 28 kg/ha of both N and
P are applied and incorporated by disking. The remaining nitrogen is surface
applied in June. Atrazine is surface applied at planting time at the rate of
3.36 kg/ha. Planting is assumed to occur on May 1 each year. The second
nitrogen application is assumed to occur on June 11. These same dates and
rates for applying nutrients and pesticide are. used in both management practi-
ces. Table 11-45 gives dates of tillage operations for MP1 and dates and rates
for applying fertilizer and pesticides.

Table 11-45. — Tillage operations and applications of fertilizer and pesticide
on the Georgia Piedmont field!/

Date

Field
operation

Fertilizer
N P

Pesticide

Name

Rate

74105
74122
74150
74162
74165
74274

Moldboard plow

Disk /pi ant /fertilize

Cultivate

Fertil ize

Cultivate

Harvest/shred stalks

(kg/ha)
Atrazine 3.36

1/ Management practice MP1 (continuous corn with conventional tillage).
Operations are assumed to be the same each year of simulation.
2/ 74105 is Julian day 105 in calendar year 1974.

Mississippi Delta

The Mississippi Delta farmland has flat slopes on poorly drained soils
with a relatively high water table that fluctuates considerably during the
growing season. Farmers in the Delta is form land to obtain a uniform field
slope. Rows are run in the direction of the slope to provide good drainage
along the furrows. A field drain provides drainage from the ends of the rows.
This drain is relatively broad and flat, normally is grassed, and is used as a
turnrow for farm equipment. Figure 11-46 shows a typical field in the Missis-
sippi Delta. The field is roughly rectangular in shape, with a 32-acre drain-
age area. Row length is 1,300 ft on a 0.4% grade. The rows drain directly
into a triangular-shaped channel that has a longitudinal slope of 0.1%,channel
side slopes of 10:1, and a bermuda grass cover. The drainage channel flows
through a culvert into a larger drainage ditch. Backwater occurs at the cul-
vert entrance, but free outfall occurs at the culvert outlet. The soils in the
field are Commerce silt loam. A phreatic water table in the Mississippi Delta
fluctuates from year to year and within the year, depending upon rainfall
amounts and time of occurrence. These fluctuations cause rooting depths to
vary from year to year. A maximum rooting depth of 40 in was estimated +

332

represent normal or average conditions

Figure 11-46 also shows a typical row cross-section. In this figure over-
land flow is assumed to occur on the row ridges and concentrated flow in the
furrow is assumed to be channel 1 for the erosion model. Channel 2 for the
erosion model is the field drain.

Typical field operations consist
of several diskings between cotton
harvest in the fall and seedbed pre-
paration the following spring. Sev-
eral cultivations are made during the
cotton-growing season. These field
operations are shown in table 11-46
for the conventional management prac-
tice (MP1) along with fertilization
and pesticide applications. A spec-
trum of pesticides is used to control
weeds and insects in the clean-till
system.

Few optional management prac-
tices are practical, acceptable by
the farmer, and effective for con-
trolling erosion in the Delta. The
second management system (MP2) consi-
ders maximum erosion control under
the clean-till cotton production.
This system includes ryegrass as a
winter cover crop. Ryegrass would be
seeded after cotton is harvested to
provide protection from erosion dur-
ing the winter. It would be disked
before preparing the seedbed and applying herbicides and fertilizer. Field
operations for management practice MP2 are shown in table 11-47.

TYPICAL ROW

FURROW= CHANNEL No. I

FIELD DRAIN = CHANNEL No. 2

Figure 11-46. — Representative field,
Mississippe Delta.

Western Tennessee

The western Tennessee area considered for application has by steep slopes
with severely eroded soils. Erosion rates are high with continuous row crop-
ping. Considerable interrill and rill erosion occurs on the steep slopes with
deposition on the toe of the slope or in the concentrated flow where slopes are
much flatter. Most of the land is class V because of slopes and erosion and
normally would not be recommended for farming by the Soil Conservation Service.
Terraces constructed on these slopes would be about 50 ft apart and would be
objectionable to the farmer. This site represents a special problem and demon-
strates what could be expected under two extreme management practices: (1)
continuous corn with conventional tillage and (2) permanent fescue harvested
annually for seed (a cash crop for possible replacement of corn). Complete

2/ Personal communication with G. H. Willis
:aton Rough, La.

soil scientist, USDA-SEA-AR

333

Table 11-46.-

Date

-Field operations for Mississippi Delta management practice 1 for
1974

Field operation

Fertilizer
N P

Pesticide

Name

Rate

74035^

Disk

74042

Disk/herbicide

74064

Disk/bed/fertilize

74109

Rebed/knock down/plant/herbicide

74127

Cultivate

74133

Cultivate

74143

Cultivate

74149

Herbicide

74154

Herbicide

74165

Cultivate

74172

Fert i 1 i ze/herbi ci de

74182

Cul ti vate/herbi ci de

74198

Insecticide

74203

Insecticide

74210

Insecticide

74218

Insecticide

74225

Insecticide

74232

Insecticide

74239

Insecticide

74247

Insecticide

74253

Insecticide

74260

Insecticide

74320

Harvest

74324

Cut and shred stalks

74325

Disk

(kg/ha) (kg/ha!

100

90,

(kg/ha!

Fluometuron 1.5

Trifluralin 1.0

MSMA

Diuron

Methyl Para-
thion/EPN

.5/
.5

Methyl Para-
thion/EPN

.5/
.5

Methyl Para-
thion/EPN

.5/
.5

Methyl Para-
thion/EPN

.5/
.5

Methyl Para-
thion/EPN

.5/
.5

Methyl Para-
thion/EPN

.5/
.5

Methyl Para-
thion/EPN

.5/
.5

Methyl Para-
thion/EPN

.5/
.5

Methyl Para-
thion/EPN

.5/
.5

Methyl Para-
thion/EPN

.5/
.5

1/ 74035 is Julian day 35 in calendar year 1974,

334

Table 11-47. — Field operations for

Mississippi Delta management practice 2 for
1974

Date

Field operation

Fertilizer
N P

Pesticide

Name

Rate

74078^

74107

Disk

Disk/bed/knock down/fertilize/
Plant cotton/herbicide

(kg/ha) (kg/ha

100

74127

Cultivate

74143

Cultivate

74149

Herbicide

74154

Herbicide

74165

Cultivate

74172

Fert i 1 i ze/herbi ci de

74182

Herbicide

74198

Insecticide

74203

Insecticide

74210

Insecticide

74218

Insecticide

74225

Insecticide

74232

Insecticide

74239

Insecticide

74247

Insecticide

74253

Insecticide

74260

Insecticide

74293

Harvest/shred stalks

74294

Disk/plant ryegrass

(kg/ha)

Trifluralin/ 1.0
Fluometuron 1.5

MSMA
MSMA

MSMA
Diuron

Methyl Para-
thion/EPN

.5/
.5

Methyl Para-
thion/EPN

.5/
.5

Methyl Para-
thion/EPN

.5/
.5

Methyl Para-
thion/EPN

.5/
.5

Methyl Para-
thion/EPN

.5/
.5

Methyl Para-
thion/EPN

.5/
.5

Methyl Para-
thion/EPN

.5/
.5

Methyl Para-
thion/EPM

.5/
.5

Methyl Para-
thion/EPN

.5/
.5

Methyl Para-
thion/EPN

.5/
.5

1/ 74078 is Julian day 78 in calendar year 1974,

335

information is unavailable for the western Tennessee site.

Figure 11-47 is a soils and drainage map of the selected field. This
field is long and slender, which significantly attenuates runoff peak rate at
the field outlet. Soil on the hilltops and steep slopes is Loring silt loam
with a hardpan approximately 26 in below the surface. Drainage from the hill-
sides concentrates in the alluvial Collins silt loam, which has a slope of 0 to
2%. Since a topographic map is unavailable, the soils map was used to generate
the representative overland flow profile (fig. 11-48). The average slope from
the soils map represented each slope segment, that is, a "B" slope ranges from
2 to 5%. Thus, the average 3.5% was used. A "D" slope ranges from 8 to 12%,
and the average 10% was used. The Collins silt loam in the alluvial valley has
a 0 to 2% slope. The side slope from the toe of the Loring was assumed to have
a 1% slope, whereas the slope in the direction of concentrated flow is assumed
as 2%.

0 400 800

SCALE IN FEET

LOBS

L0D3

Z-L0B3

LORING SILT LOAM, 2-5% SLOPE, SEVERELY ERODED

LORING SILT LOAM, 8-12% SLOPE, SEVERELY ERODED

COLLINS SILT LOAM, 0-2% SLOPE

FIELD BOUNDARY

SOIL /SLOPE /EROSION BOUNDARY

CONCENTRATED FLOW

Figure 11-47. — Soils-drainage map of western Tennessee field,

The lack of a topographic map will cause inaccurate estimates of runoff
and erosion even if the model was exact. This is of little consequence since
the model is for comparing management practices. Absolute accuracy is not as
important as the relative magnitudes between practices.

Table 11-48 shows field operations and dates of herbicide and fertilizer
application for management practice 1. These operations are normal for the
area with normal planting and harvest dates. For the second management prac-
tice (MP2) with permanent fescue grass, the fescue is assumed to be established
at the beginning of simulation. This assumption is in keeping with the previ-
ous description of extreme conditions for hydrology and erosion.

3/ Data on soils and general information were abstracted by SCS personnel
from a University of tennessee master's thesis.

336

FIELD EDGE

LORING
10% SLOPE

300 200

RELATIVE DISTANCE, FEET

Figure 11-48. — Representative overland flow profile estimated from soil map

and average slopes.

Table 11-48. — Field operations for western Tennessee, management practice 1,
continuous corn with conventional tillage

Date

Field Operation

Fertilizer
N P

Pesticide

Name

Rate

74092
74119

74122

74154
74177
74198
74309

Moldboard

Disk

Disk/plant/fertilize/

apply herbicide.
Cultivate
Cultivate
Cultivate
Harvest/shred

(kg/ha) (kg/ha)

140

Atrazine 3.36

1/ 74092 is Julian day 92 in calendar year 1974.

APPLICATIONS AND PARAMETER ESTIMATES

The rest of this chapter is presented by components. That is, hydrology
is presented for the applications in one section so the user can see these
applications without having to read through erosion, plant nutrients, and pes-
ticides. All erosion applications are in a single section so the user can see
the different conditions represented. Parameter values are discussed suffi-
ciently to help the user understand the estimation process.

337

HYDROLOGY

Application of both hydrology options in this chapter will help the user
select parameter values. Hydrology option 1 is given for both management
alternates on the western Tennessee site. Hydrology option 2 is used for the
Georgia Piedmont and Mississippi Delta.

Hydrology Option One

Western Tennessee

The process for obtaining parameter values is designed to be as objective
as possible while maintaining general applicability. Judgment is required,
however, in estimating missing data, crop production, soil characteristic vari-
ations, and so forth. To demonstrate this process, each parameter estimate is
decribed in detail for management practice 1. Only parameter estimates that
change values in converting from MP1 to MP2 are presented for MP2.

Management Practice 1
value.

Card 1 to 3:

The source of information is given for each parameter

TITLE = DAILY HYDROLOGY PARAMETERS - WESTERN TENNESSEE
MANAGEMENT PRACTICE 1
CONTINUOUS CORN - CONVENTIONAL TILLAGE (HYDONE)

Alphanumeric information describing the problem.

Card 4:

BDATE = 74001 Beginning date of simulation {year and Julian day).

FLGOUT = 1 Specifies output for each storm and for annual summaries.

FLGPAS = 1 Creates a hydrology file for use by the erosion model.

FLGOPT = 1 Indicates that the hydrology option 1 model will be used.

Card 5:

DACRE = 69.2 Drainage area of the field measured from drainage map,

acres.

RC = 0.10 Effective saturated conductivity of the predominant soil,

Loring silt loam, in/hr, obtained from description of
soils given by Hoi tan and others (I).

FUL = 0.80 Fraction of plant -avai lable water storage filled at field

capacity. FUL = (Field capacity - BR15)/(Porosity - BR15)
where field capacity is estimated as the volumetric con-
tent at 0.1 bar tension and BR15 is the volumetric water

338

content at 15 bars tension obtained from Hoi tan and others
(I).

BST = 0.5 Fraction of plant-available water storage filled when sim-

ulation begins. Since this simulation begins on January
1, the plant-available water storage is assumed to be half
full. If the simulation began in the fall after harvest,
BST would be estimated much lower (0.5 - 0.1) because the
crop normally uses most water in the root zone.

CONA = 3.5 Soil evaporation parameter. A value of 3.5 is satisfac-

tory for most soils. The suggested range is from 3.3 for
sands to 4.5 for loams.

P0R0S = 0.48 Soil porosity in the root zone, cm-Vcm^, obtained from

Holtan and others ( I) .

Card 6:

SIA = 0.2 Initial abstraction coefficient for SCS runoff equation.
The 0.2 value generally is recommended.

CN2 = 86 Two condition SCS curve number (5_, tables 7.1 and 9.1).

CHS = 0.01 Channel slope, ft/ft, determined by dividing the elevation
difference by the distance along the main drainageway from
the field outlet to the most distant point.

WLW =5.4 Field length-width ratio determined by dividing the square
of the channel length as defined in CHS by the drainage
area. For this calculation, channel length is expressed
in miles and drainage area in square miles.

RD = 27 Root depth, inches, estimated as the depth to the hardpan.
If no hardpan exists, the root depth for most crops is
about 36 in. This depends on the type of crop, soil, and
production level .

Card 7:

UL(l-7) = 0.27, 1.15, 1.44, 1.28, 1.03, 0.93

Plant -avai lable soil water storage, inches, for each of
seven storages. UL = Porosity - (BR15) (RD)(D) , D = 1/36
for top storage, 5/36 for second storage, and 1/6 for 5
lower storages, obtained from Holtan and others {I).

Cards 8, 9:

TEMP(1-12) = 41.2, 46.0, 53.5, 60.4, 68.5, 74.6, 79.4, 77.7, 68.6, 59.6, 48.2,
40.6 Average monthly temperature at Nashville, Tenn., obtained
from the Climatic Atlas of the United States (6).

339

Cards 10, 11:

RAD(1-12) = 149, 228, 322, 432, 503, 551, 473, 403, 308, 208, 150.
Average monthly solar radiation, langleys/day, at Nash-
ville, Tenn., obtained from the Climatic Atlas of the
United States (6) or table 1 1-7 .

Card 12:

GR = 1.0 Winter cover factor. Essentially no winter cover is as-

sumed for continuous corn with conventional tillage.

Cards 13 to 25: Number of cards varies for different crops.

0.0

132

0.0

167

0.09

182

0.19

196

0.23

211

0.49

225

1.16

240

2.97

254

3.00

269

2.72

283

1.83

309

0.00

366

0.00

Julian day and leaf area index for the crop grown during the first year,
or simulation (table II-8).

Rainfall data—Daily rainfall data were used from USDA-SEA-AR records near
Clarksdale, Miss. One year's data are placed on 37 cards.

Since management practice 1 specifies continuous corn, the leaf area index
and winter cover factor remain the same thoughout the simulation. New values
of these variables must be input each year if the crop changes.

Management practice 2--0nly three input changes are required in converting from
management practice 1 to management practice 2. Since these parameters (CN2,
GR, and LAI) are the primary indicators of management changes, adjusting other
parameters usually is unnecessary when considering management practices.

The two condition SCS curve number is 79 (5^, tables 7.1 and 9.1). The
winter cover factor is GR = 0.5 because the fescue grass should provide an
excellent winter cover.

Leaf area index data are unavailable for fescue grass. Information was
obtained for growing period, approximate dates and rates of fertilization,
dates of harvest, approximate yields, and recommended herbicide practices.
Fescue grass is a cool -season grass that begins rapid growth about mid-February
in western Tennessee. A balanced fertilizer is applied at that time at the
rate of 60 lb/acre nitrogen, 20 lb/acre phosphorus, and 30 lb/acre potassium.

340

A light application of 2,4-D is made in late April or early May at the rate of
1.5 kg/ha. Growth rate begins to decrease the beginning of April and reaches a
maximum by mid-May. Seed is ready for harvest by late June. Seed is harvested
with a grain combine that leaves much plant material standing. A good yield is
1,000 lb/acre of seed and 5,000 lb/acre of dry matter. Fescue does not go com-
pletely dormant during the summer, but it grows little until temperatures drop
in mid-September. The growth rate again declines by the first of November but,
like winter small grain, a transpiring canopy exists throughout the winter.
Table 11-49 gives the leaf area index used in the hydrology model.

Table 11-49. — Leaf area index for fescue grass, management practice 2 (MP2),

western Tennessee-/

Date

Julian day

Leaf Area Index

1_1 001

2-15 046

3_1 091

5-15 135

6-15 166

7_1 182

7_2 183

9_15 258

11_1 305

12-31 366

0.35

.40

2.10

2.80

2.60

.20

.25

.35

1/ Personal communication with S. R. Wilkinson, USDA-SEA-AR, Southern
Piedmont Conservation Research Center, Watkinsville, Ga.

Interpretation of Results

Table 11-50 compares results for the two simulations. Although the simu-
lation period was only 3 yr, a wide variation in hydrologic conditions was
observed. For example,, rainfall ranged from 34.5 to 70.7 in. Average annual
values in table 11-50 shows that management practice 2 gives less surface run-
off and more percolation than management practice 1. These results seem rea-
sonable because the fescue grass increases the infiltration rate of soil consi-
derably. Since evapotranspi ration is essentially the same for the two manage-
ment practices, percolation must be higher if infiltration is increased.

Hydrology Option 2

Many parameters necessary for hydrology option 2 are the same as for
option 1. The two example applications show how parameters are selected where
judgment is reflected. Parameters are discussed as listed in table 1 1-5.

Georgia Piedmont Management Practice 1

The measured area (DACRE) of the field/watershed in figure 11-44 is 3.2

341

Table 11-50. — Results from hydrology option 1 for management practices 1 and 2

in western Tennessee

Rainfall

Management 1

>ractice 1

Management Practice 2
d..™-p-f Evapotrans- Pen
Runoff piration lati

Year

Runoff

Evapotrans-
piration

Perco-
lation

:o-
ion

- -(in)-

1974

.71

.17

.72

8.61

16.04

.45

.98

1975

.90

14,

.09

.24

7.68

9.04

.74

.15

1976

.53

.77

.07

4.94

4.41

.99

.66

Average
Annual

.05

.01

.68

7.08

9.83

.06

.26

acres. RC is set at 0.19 in/hr based on information on soils ( lj , SCS
hydrologic soil group B ( 5_) , and guidelines in table II-9. A better procedure
is to estimate RC and GA by best fit with infi Urometer measurements, but these
are often unavailable. Management practice 2, with a grassed waterway, will
have higher RC.

Volumetric saturation at field capacity, FUL, is estimated as 0.75. FUL
will be higher for clay loams and lower for sandy soils. A corresponding value
for available water content at the beginning of simulation, BST, is taken as
0.50, which is unknown, but simulation is not sensitive to this starting value.
BST = 0.5 assumes that the soil contains half its capacity of stored water at
the beginning of the year.

The soil evaporation parameter, C0NA, is set at 3.75 according to guide-
lines in CREAMS, volume I, chapter 2 (3). Total porosity, P0R0S, is 0.41 as
taken from data in Holtan and others (I). This value is a rough average of
porosities reported for upper soil layer samples.

BR15, representing volume of immobile water, ranges from about 5 to 30%,
depending on soil type and conditions. Measured values in Holtan and others
{I) vary greatly from layer to layer but are generally low for sandy soils and
high for clay soils. A mean value is 0.17 in/in for the upper layers of the
Cecil soil used here.

Average monthly values of temperature and radiation are in the Climatic
Atlas (6) (or CREAMS, vol. II, table II-7). Since this is a cropped watershed,
winter cover is small and GR = 1.0. The leaf area index values for the corn
crop are available from table 1 1-8.

The depth of surface soil layer, DS, arbitrarily is taken as 2.0 in, and
the root layer depth, DP, is assumed to be 22 in, making rooting depth, RD,
equal to 24 in.

The infiltration parameter, GA, ordinarily is found for this soil (group
B) from table II-9. The value used (13 in) is higher than the recommended
range for a hydrologic group B soil but was chosen from comparisons to actual
infiltrometer data. Tilled soils exhibit higher values of GA than untilled or

342

undisturbed soils, which should be reflected when choosing a value for GA.

The remaining parameters reflect the hydraulic and topographic conditions
governing overland flow and, in this model, the estimation of peak runoff
rates. The Manning roughness value (RMN) of 0.03 is typical of flow along
plowed furrows. Values for slope and distance of a "typical" overland flow
path should reflect actual flow paths to make best estimates of runoff peaks.
Flow often will follow furrows rather than the topographic "downhill" direc-
tion, which modifies slopes and lengths of flow.

The watershed in figure 11-44 exhibits overland flow along furrows until
the bottom of the swale is reached. Flow should move in a broad, rough channel
to the outlet. Furrows in this example run east and west. The "channel" flow
should be fairly rapid, and measured mean "furrow" distance is about 250 ft. A
total estimated (weighted) flow length, XLP, is taken as 350 ft. The effective
(weighted) field slope is estimated as 0.015 (1.5%). This value should put
most weight on flow along furrows where the runoff water spends most of its
time. The formal procedure for getting optimum mean slope (CREAMS, vol. I, ch.
2, eq. 1-36) is not used since it applies to cascaded overland flow planes
rather than a combination of planes and channels. In this example, furrow
slope and swale slope are nearly equal.

Interpretation of Results

Results from hydrology option 2 consist of a summary of input parameters,
a table of daily values of rainfall, runoff, and other water-balance informa-
tion for all days on which rainfall occurs, plus the passfile created for sub-
sequent model component simulations.

Simulation results for the Georgia Piedmont application are summarized in
figure 11-49. The fallow saturated-hydrologic conductivity in the program is
0.8 times the normal (cultivated) conductivity.

Results show that in each year total evapotranspiration is a large part of
rainfall. For the lowest rainfall year, 1973, soil water never became large
enough to cause deep percolation. Although it is unlikely that no soil water
moved below the root zone, percolation was very small. The nature of the mod-
el's approximation to soil -water movement and simplification of the soil water
system will insure that years with low percolation cannot be simulated accu-
rately.

Mississippi Delta Management Practice 1

The Mississippi Delta is topographically much more straightforward to
represent since it is rectangular and, therefore, has a uniform flow length for
flow along the furrows. The soil parameters were obtained from Lund and Loftin
( 2J . Table 1 1 -9 shows that the soil is in a low B or a high C group, with
hydraulic conductivity (RC) of 0.16. A value of 11 for GA is consistent with
this classification.

Parameters FUL, BST, P0R0S, and BR15 were assigned from values typical for

343

HYDROLOGY SUMMARY

DAILY HYDROLOGY PARAMETERS - GEORGIA PIEDMONT

MANAGEMENT PRACTICE ONE

CONTINOUS CORN - CONUENTIONAL TILLAGE

1374

MONTH

JAN
FEB
MAR
APR
MAY
JUN
JUL
AUG
SEP
OCT
NOU
DEC

TOT

RAIN

2.700
4.110
1.920
2. GOO
5.420
5.230
4.150
5.780
1.850
0.3G0
1.1G0
4.320

40.260

RUNOFF

PERC

AUG SW

0.008

2.157

0.43S

2.270

0.008

1.618

0.173

1.331

0.G44

3.374

1.254

3.153

0.58G

4.726

0.132

6.440

0.000

2.133

0.000

0.554

0.000

0.782

0.203

1.666

3.51G

30.323

1375

0.000

2.514

1.380

2.830

0.106

2.544

0.763

2.735

0.371

2.742

1.257

2.351

0.000

0.901

0.000

0.353

0.000

0.457

0.000

0.095

0.000

0.282

0.433

1.484

4.316

.653

MONTH

RAIN

RUNOFF

PERC

AUG SU

JAN

5.020

0.430

2.321

2.522

2.977

FEB

7.170

1.035

2.460

3.611

2.920

MAR

3.730

2.361

2.330

3.682

2.864

APR

3.330

0 . 334

2.173

1.260

2.717

MAY

G.070

0.633

3.123

1.686

2.771

JUN

3.550

1.071

2.336

1.067

2.473

JUL

4.670

0.043

6.012

0.000

0.615

AUG

2.340

0.000

2.134

0.000

0.165

SEP

5.370

0.333

2.384

0.000

1.200

OCT

0.350

0.000

0.316

0.000

2.157

NOU

0.000

0.323

0.000

1.721

DEC

0.000

0.262

0.000

1.431

TOT

48.250

7.505
ANNUAL

28.602
AUERAGES

13.828

2.001

MONTH

RUNOFF

PERC

3.372

sue

JAN

3.8S0

0.243

2.233

1.261

2.745

FEB

5.640

0.735

2.365

2.496

2.875

MAR

5.850

1.485

2.274

1.834

2.704

APR

3.265

0.556

2.032

1.014

2.726

MAY

5.745

0.633

3.243

1.323

2.r"5G

JUN

4.420

1.163

3.074

1.162

2.415

JUL

4.410

0.314

5.363

0.000

0.758

AUG

4.060

0.036

4.317

0.000

0.567

SEP

3.610

0.170

2.532

0.000

0.828

OCT

0.355

0.000

0.635

0.000

1.126

NOU

0.580

0.000

0.555

0.000

1.001

DEC

2.460

0.105

0.364

0.217

1 . 458

1.830

TOT 44.255 5.511 23.765

Figure 11-49. — Annual summaries of simulation results using hydrology option 2,
Ga. Piedmont application, MP1.

344

soils of this type since measurements were unavailable. Although normal root-
ing depth for cotton can extend to 4 ft, 40 in reflect the expected effect of
the fluctuating high water table in this area. The winter cover factor, GR, of
0.50 reflects the practice of leaving cotton plant residue on the soil over
winter.

The relatively large row channels indicated a somewhat lower roughness
value than ordinary furrows, and Manning's n (RMN) was chosen as 0.02. The ef-
fective slope of 0.003 is a weighted value combining the row slope of 0.004 and
the channel slope of 0.001. Effect of the intercepting channel was included in
the overall effective flow length of 2,000 ft, especially since it is a grass-
lined channel. The value of XLP is, nevertheless, somewhat smaller than the
total of the row and channel lengths in figure 11-46.

Temperature and radiation data were from the nearest U.S. National Weather
Service recording stations. In this situation, radiation data were from
Shrevesport, La., and temperature data were from Jackson, Miss. Leaf area
index data for cotton were taken from table 1 1-8.

Interpretation of Results

Figure 11-50 shows summary output information from the hydrology subrou-
tines for HYDTW0 for this example. Compared to the Georgia Piedmont example,
total evapotranspiration is a lower percentage of the total rainfall but still
is larger than the corresponding ET in Georgia. Rainfall here is much larger,
although the 3 yr represent both a year with lower and higher than average
rainfall. The lower RC intuitively causes a higher proportion of runoff.
Since with the number of days of high soil-water content is large, percolation
is also relatively large.

Not shown in either output HYDTW0 data example is the predicted sequence
of peak flows for the runoff events. These naturally will reflect the values
chosen for surface response-related parameters, including RMN, SLOPE, and XLP.
Management practices strongly are reflected in RMN and XLP values, and peak
flows are affected considerably as a result.

EROSION

The western Tennessee and Mississippi Delta examples illustrate selection
of parameter values and application of the erosion component of the model. The
western Tennessee site is discussed first because it is typical of many culti-
vated fields. The Delta site shows a special application of the model.

Only the most significant cards and parameters are discussed. Refer to
chapter 2 for card sequence and identification and definition of parameters.
Several of the first cards of each input file are shown in accompanying fig-
ures.

Western Tennessee
Figure 11-51 shows the initial part of the parameter file. Major entries

345

HYDROLOGY SUMMARY

BREAKPOINT HYDROLOGY PARAMETERS - MISSISSIPPI DELTA

MANAGEMENT PRACTICE ONE

CONTINUOUS COTTON - CONUENTIONAL TILLAGE

1374

MONTH

RAIN

RUNOFF

PERC

AUG SU

JAN

8.880

0.312

1.25G

4.G72

3.304

FEB

3.470

0.400

1.152

2.173

4.158

MAR

1.320

0.033

1.082

0.G81

4.038

APR

5.380

0.872

1.328

2.385

4.084

MAY

14.620

2.307

3.073

8.533

4.148

JUN

3.050

2.G07

G.4G2

2.477

3.704

JUL

5.870

2.280

5.084

0.000

0.G5G

AUG

8.010

1.S05

3.504

0.000

O.G30

SEP

3.G50

0.437

5.GGG

0.000

1.818

OCT

1.370

0.031

0.325

0.000

0.773

NOU

4.140

0.735

1.507

0.000

3.008

DEC

4.G10

0.387

1.150

2.422

4.123

TOT

71.570

13.334

32.183

23.350

2.32G

MONTH

RAIN

RUNOFF

1375

PERC

AUG SU

JAN

4. ISO

0.223

1.128

2.720

4.22G

FEB

4.870

0.G52

1.470

3.008

4.274

MAR

7.170

0.803

1.842

4.433

4.137

APR

4.770

0.441

2.083

2.037

4.152

MAY

4.330

0.031

2.355

2.321

4.114

JUN

4.750

0.320

G.287

0.805

3.340

JUL

2.8G0

0.3G7

3.G87

0.000

0.410

AUG

5.7G0

1.G08

4.201

0.000

0.G38

SEP

4.010

0.231

3.331

0.000

0.GG5

OCT

2.230

0.243

0.8G7

0.000

0.782

NOU

3.140

0.744

1.407

4.1GG

4.144

DEC

2.G30

0.000

1.18G

1.G77

4.374

TOT

57.340

5.734

30.503

21.283

2.343

Figure 11-50. — Summary results of simulation for the Miss. Delta MP1 using

hydrology option 2.

346

197G

MONTH RAIN RUNOFF ET PERC AUG SW

JAN

3.410

0.371

1.22G

1.835

4.244

FEB

5.830

1.484

1.199

3.245

4.154

MAR

G.880

0.417

1.92G

4.228

4.230

APR

1.220

0.022

1.277

0.233

4.082

MAY

1.490

0.000

1.458

0.2G1

3.973

JUN

4.800

0.149

5.GG1

0.000

3.280

JUL

1.220

0.0G2

3.GGG

0.000

0.GG4

AUG

0.140

0.000

0.285

0.000

0.05G

SEP

3.400

0.352

2.885

0.000

0.470

OCT

3.140

0.000

1.614

0.000

0.551

NOU

1.900

0.000

1.08G

0.000

1.919

DEC

1.380

-0.000

1.024

0.000

2.937

TOT 34.810 2.857 23.30G 9.803 2.547

ANNUAL AUERAGES

MONTH

RAIN

RUNOFF

PERC

AUG SW

JAN

5.483

0.502

1.203

3.07G

4.125

FEB

4.723

0.84G

1.273

2.809

4.195

MAR

5.323

0.421

1.617

3.134

4.175

APR

3.790

0.445

1.5G5

1.772

4.10G

MAY

7.033

0.979

2.495

3.707

4.078

JUN

G.200

1.025

G.137

1.094

3.442

JUL

3.317

0.903

4.14G

0.000

0.577

AUG

4.G37

1.071

2.GG3

0.000

0.441

SEP

3.G87

0.380

3.980

0.000

0.984

OCT

2.447

0.113

1.135

0.000

0.704

NOU

5.0G0

0.493

1.334

1.389

3.024

DEC

2.873

0.129

1.120

1.36G

3.813

TOT 54.573 7.309 28.GG8 18.347 2.805

Figure 11-50. — Summary results of simulation for the Miss. Delta MP1 using
hydrology option 2--continued.

347

CARD

EROSION PARAMETER DATA

EROSION PARAMETERS - WESTERN TENNESSEE

MANAGEMENT PRACTICE ONE

CONT Li

74000

0 4

0.000

,000

,000 0.000

0.000

0.150

0.G50

,200

020 0.000

0.000 0.000

0.000

G9.210

276.000

,074

,030 0.120

0.010 61.000

18.400 174.000 4.84

1.000

0.400

2 1

20.000

10.000

,035

,010 2.000

2.000 0.000

5G8.000

G.500

,800

20,

,000

0.000

0.010

284,

.000

,020 5G8.000

0.030

4 1

20.000

.035

,005 31.000

2.000 0.000

4900.000

G9.210

.800

20,

,000

0.000

0.0051380

.000

,0061620.000

0.0102400.000

0.0153000.000 0.01!

74001

74045

1.000

0.200

1.000

0.800

1.000

0.030

1 1

0.000

0.0G0

0.000

0.400

0.000

100.000

0.000

0.330

0.000

0.330

•

0.000

10.000

1 1

0.000

O.OGO

0.000

0.400

0.000

100.000

0.000

0.330

0.000

0.330

0.000

20.000

7404G.'

74092

1.000

0.250

0 0

74093

74119

1.000

0.430

1.000

0.040

1 1

0.000

0.045

0.000

0.150

0.000

0.330

0.000

0.330

1 1

0.000

0.045

0.000

0.100

0.000

0.330

0.000

0.330

Figure 11-51. — Partial parameter file for the erosion component with
application to western Tennessee, management practice 1.

348

by cards will be discussed.

The alphanumeric information identifies such items as site

Cards 1 to 3:
management practice

and other important identification factors.

Card 4: BDATE = 74000. BDATE can be greater than the first PDATE in the
parameter file, but it must be less than the first storm date, SDATE, in the
hydrology file. FLGPRT = 0 so that the model will use the primary particle
distribution of the soil to compute the sediment particle specifications from
internal relationships in the model. FLGSEQ = 4 to accommodate a watershed
with two stream orders (fig. 11-52).

SUBWATERSHED DIVIDE

OVERLAND FLOW PATH

SECONDARY FLOW
CONCENTRATION

FLOW CONCENTRATION

0 400 eoo

SCALE IN FEET

Figure 11-52. — Field, subwatershed, channel, and overland flow definitions.

Card 5: The defaults for variables on this card are used by leaving this
card blank.

Card 6: The primary particle size and organic matter contents were esti-
mated from SCS soil survey information. Specific surface area variables were
left blank so that default values are used since better information was una-
vailable.

Cards 7 to 8: These cards are absent because default sediment particle
specifications calculated within the model from primary particle size data are
used. If FLGPRT = 3 on card 4, cards 7 and 8 must be present.

Card 9: This watershed is made up of several small subwatersheds. Aver-
age values for topographic factors must be estimated, or representative values
must be chosen. Typical subwatersheds are shown in figure 11-52. Since a con-
tour map was unavailable, a typical overland flow profile was constructed from
soil survey maps (fig. 11-53). The parameter values are DATOV = 69.21 acres,
which is the area of the total watershed. An average value for the subwater-
sheds could have been used. The only difference in the output would have been
in total amount (lb) of sediment produced on the overland flow areas. DATOV
does not affect the sediment yield per unit area and concentration of sediment
in the runoff. SLNGTH, AVGSLP, SB, SM, SE , XIN(3), YIM(3), and XIN(4), YIN(4)

349

FIELD EDGE

500

400 300 200

RELATIVE DISTANCE, FEET

100

Figure 11-53. — Representative overland flow profile constructed from soils^

slope data.

are, respectively, 276 ft, 0.074,
and 4.84 ft.

0.030, 0.120, 0.010, 61 ft, 18.4 ft, 174 ft

Cards 10 to 11: A soil crodibility of 0.4 ton/acre/EI was used for all
locations along the representative overland flow profile. Values for soil
erodibility are available from SCS.

Card 12: NS = 3: Three segments were used to describe the channel pro-
file. FLAGC = 1 selected a triangular channel. FLAGS = 1 specified that the
program use the equations for spatially varied flow. CONTL = 2 set uniform
flow as control at the channel outlet. SECTN = 1 selected a triangular channel
for the outlet control section.

Card 13: SIDSLP = 20 for the side slope, BOTWID = 10 ft, OUTMAN = 0.035
for Manning's n, and OUTSLP = 0.01 for slope of the outlet control channel.
The rating parameters RA, RB , and YBASE are not used although values of 2, 2,
and 0 are entered.

Card 14: LNGTH = 568 ft was the average length for the channel in each
subwatershed. DATCH = 6.5 acres for the average drainage area in the subwater-
sheds above the channel outlet, and DAUCH = 0.8 acre for the average drainage
area above the entrance to the channel in the subwatersheds. A side slope, Z =
20, for the secondary flow concentrations was used because the concentrations
are farmed over.

350

Card 15: TX and TS are locations and slopes along the channel profile.
Distances are referenced to the channel outlet. Therefore, TX(1) = 0 at the
channel outlet. Slope along the channel was estimated from the soil survey
map.

Cards 12 to 15: These cards repeat for the main channel of the watershed.
FLAGS was set to 1 to use the energy gradeline curves to consider backwater for
an assumed restricted outlet. CONTL = 4 was used to express control by a rat-
ing curve at the outlet. The rating coefficients RA = 31, RB = 2, and YBASE =
0 were selected to give estimated depths for assumed discharges. DATCH = 69.21
acres is the total watershed area. Channel slopes were estimated from the soil
survey maps.

Cards 16 to 17: These cards are absent because no pond element is used.

Card 18: PDATE = 74000 and BDATE = 74045. These are the dates between
which the following parameter values are valid. This period is one of winter
stalk cover.

Card 19: NC, NP, and NM = 1 because of uniformity along the slope. In
the first set of updateable parameter values, a nonzero value must be assigned
to NC, NP, and NM to initialize the overland flow parameters.

Card 20: Since a uniform soil loss ratio is assumed for the slope, XCIN
(1) = 1.0. At this crop stage, the soil loss ratio for about 1-1/2 tons/acre
of surface residue is 0.20 from tables 11-20 and 11-23, and figure 11-23. This
represents an average C between standing stalks left by a cornpicker and stalks
uniformly shredded with a shredder.

Card 21: PIN(l) = 0.80 for partial contouring, which results from the
assumed "parallel to fence farming."

Card 22: Manning's n is a function of cover and roughness. A relative
smooth surface and 1-1/2 ton/acre of cornstalks give MIN of 0.03 (table 11-26).

Card 23: All variables on the card are set to 1 because of uniformity
along the channels and to initialize the parameter values.

Card 24: The first TX, that is, TX(1), is always 0 because the reference
is at the outlet end of the channel. TN = 0.06 (table 11-28) is assigned to
Manning's n. This value applies to the entire channel length.

Card 25: Critical shear stress, TCR, is set fairly high at 0.4 lb/ft2
(table 11-29) because the soil is assumed to have consolidated since the last
cultivation in the summer.

Card 26: The effect of cover breakdown is ignored by setting TCV to the
large number (100 lb/ft2), which greatly exceeds values for the flow's shear
stress.

Card 27: TON, the depth to the nonerodible layer, is an initial value the
first time it is read. A definite value for TDM is unknown except following
tillage. Therefore, simulations are best started at the time of tillage, but

351

this was inconvenient for the problem. A value of 0.33 ft approximates the
depth of secondary tillage, the limiting depth for many fields.

Card 28: The nonerodible layer is assumed to follow the curvature of the
surface soil, that is, it is parallel to the soil surface. Therefore, IDS is
set equal to TDN.

Card 29: Although the channel is triangular, a width TW = 10 ft is speci-
fied because the model sometimes defaults to a rectangular channel.

Cards 23 to 29: These cards are repeated to describe the second channel.

Once the storm date exceeds CDATE, the model reads a new set of updateable
parameters for the next crop stage. New values are required only for those pa-
rameters that change. The cards begin repeating at card 18.

Card 18: The new dates are 74046 and 74092. This is the end of the win-
ter crop stage period. It is included to account for further decay of residue
over the winter. The field is moldboard plowed on 74093.

Card 19: Contouring and Manning's n do not change from previous values.
New input values are not read by setting NP and NM to 0. NC = 1 indicates a
new soil loss ratio that is uniform along the slope.

Card 20: The new soil loss ratio is 0.25.

Cards 21 to 22: These are absent because NP and NM = 0.

Card 23: All values are set to zero to use previous values.

Cards 23 to 29: These do not appear because previous values are used for
all parameters on the cards.

The next crop stage follows moldboard plowing on 74093. Plowing changes
the soil loss ratio, and Manning's n for overland flow and channel flow reduces
the critical shear stress and resets the depth to nonerodible layer. The next
group of cards for 74093 and 74119 are inputs for these changes.

Interpretation of Results

The results of several management options for the West Tennessee site are
discussed in a section late in chapter 2.

Mississippi Delta

Management practice 1--The Mississippi Delta site illustrates a special appli-
cation of the model. This example is quite different because the watershed is
flat and an unusual watershed representation is used.

The field is disked and bedded several times during the year. Well-de-
fined row ridges and middles that form the flow patterns exist most of the

352

year. The representation assumed is an overland area of row side slopes, chan-
nel 1 for a representative row middle, and channel 2 for the field ditch.

Figure 11-54 partially lists the input parameters,
discussed differ significantly from the first example.

The parameter values

CARD
NO

EROSION PARAMETER DATA

2
3
4

5
G
3
10
11
12
13
14
15
12
13
14
15
18
19
20
21
22
23
24
25
26
2?
28
23
23
24
25
2G
27
23
23
18

74000
0.000
0.E00
0.090
1
1.000

20.000

1300.000

0.000

5.000

1270.000

0.000

74001

1.000

0.000
0.000
0.000
0.000
0.000
0.000

0.000
0.000
0.000
0.000
0.000
0.000
7403G

0
0.000
0.500
1.500

0.370

2
3.000
0.090
0.004

3.000

32.000

0.001

74035

1
0.540
1.000
0.050

1
0.0G0
0.200
100.000
0.330
0.330
3.000

0.150

0.700

100.000

5.000

740G4

EROSION PARAMETERS - MISSISSIPPI DELTA

MANAGEMENT PRACTICE ONE

CONTINUOUS COTTON - CONUENTIONAL TILLAGE

1 0 4

0.000 0.000 0'-050 0.000

0.300 0.012 0.000 0.000 0.000 0.000
0.200 0.200 0.200 0.200 1.500 0.000

1
0.020
0.000

1.500 0.000

2 2 1

0.030 0.004 0.000
0.000 5.000

4
0.010
5.000

0.000 0.000

15.000 0.500 0.000

Figure 11-54. — Partial listing of parameter values for the Mississippi Delta

management practice 1.

Card 4: CSEQ = 4 designates overland flow (row side slopes) -
row middle) - channel (field ditch) for the watershed representation.

channel

Card 5: The Manning's n for overland flow over bare soil is increased to
0.05 to ensure that no deposition is calculated on the row side slopes. The
transport equations are intended for longer slopes. Increasing n for bare con-
ditions increases computed transport capacity.

353

Card 9: A typical row side slope is the overland flow area. The rows are
assumed to be 3.0 ft apart, and 1.5 ft is assumed for the width of a single row
side slope, which is the length of overland flow. The overland flow area for a
single row side slope for the 1,300 ft row -length is 0.0448 acre. Since two
side slopes occur per row middle, a value of 0.090 acre is used for DAT0V to
obtain the total amount of sediment draining into the row middle. Slope
length, SLNGTH, for the overland flow area is 1.5 ft. A uniform steepness of
0.2 is assumed. Therefore, AVGSLP, SB, SM, and SE = 0.2. Since the slope is
uniform, XIN(3), YIN(3), and XIN(4), YIN(4) = 1.5, 0.0.

Card 12: Since the slope along the row middle is assumed to be uniform,
NS = 1, that is, only one slope value is required. FLAGC = 2 for a rectangular
channel in the row middle. Since FLAGS = 2 (table 11-27), the slope of the
energy gradeline (friction slope) equals the channel slope. Flow rates are
small in row middles, and backwater effects do not extend beyond a few feet up
the middles.

The parameters C0NTL and SECTN are not used since FLAGS = 2, but dummy
values are entered.

Card 13: Although values on this card are not used, the card must be pre-
sent with nonzero dummy values, except for YBASE.

Card 14: LNGTH = 1300 ft for the length of the row middles. The drainage
area DATCH for a single row is 0.090 acre. The drainage area DAUCH at the up-
per end of each middle is zero. The channel side slope Z is approximated at
5:1 (that is, 5.0).

Card 15: The slope along the row middle is 0.004. The entry on this card
is 0.0, 0.004.

Cards 12 to 15: These cards repeat for the field ditch and are similar to
other channel cards, except that backwater and a rating curve are assumed for
outlet control.

Card 25: TCR for the second channel was set to 0.70 lb/ft2, a relative-
ly large critical shear stress representing the long period of consolidation
since tillage.

Cards 27 to 28: TDN and TDS were set to large value, 100 ft, to ignore
the effect of the nonerodible boundary, which was assumed not to exist.

Management Practice 2 - This management practice included a reduced number of
tillage operations, winter cover, and a 20-ft grass strip at the end of the
rows. Figure 11-55 shows the first part of the input file for the problem.

Skip to card 23.

Card 23: The 20-ft grass strip at the end of the rows requires a differ-
ent Manning's n, critical shear stress, and channel width from that used for
the cultivated portion of the rows. Therefore, NN = 2, NCR = 2, NCV = 1, NDN =
1, NDS = 1, and NW = 2.

Card 24: A Manning's n of 0.10 is assumed for the grass and 0.06 for the
cover crop, which begins 20 ft up the row. Entries on the card are 0.0, 0.15
(grass strip), 20.0, and 0.06 (tilled part of row).

354

CARD

ER05I0N PARAMETER DATA

ER05I0N PAPAMETER5 -

MI53ISSIPPI

DELTA

MANAGEMENT PRACTICE TWO

C0NTINU0U5

COTTON -

MODIFIED TILLAGE -

WINTER COUER

74000

0.000

,0 50

0.000

0.200

0.500

0.300

0.012

,000

0.000 0,

.000 0.000

0.090

1.500

0.200

,200

0.200 1,

.500 0.000

1.000

0.370

20.000

3.000

0.030

0.004

,000

0.000 0,

.000

1300.000

0.090

0.000

5.000

0.000

0.004

5.000

3.000

0.020

0.010

15.

.000

0.500 0

.000

1270.000

32.000

0.000

5.000

0.000

0.001

74001

7404G

1.000

0.200

1.000

0.050

0.000

0.150

20.000

0.0G0

0.000

0.700

20.000

0.300

0.000

100.000

0.000

0.330

0.000

0.330

0.000

3.000

20.000

0.500

0.000

0.150

0.000

0.700

0.000

100.000

0.000

100.000

0.000

100.000

0.000

5.000

74047

74078

1.000

0.150

0.000

0.150

20.000

0.100

0.000

0.700

20.000

0.400

74079

74107

1.000

0.580

0.000

0.150

20.000

0.040

0.000

0.700

20.000

0.200

0.000

0.330

0.000

0.330

0.000

3.000

20.000

0.500

1.500 0.000

Figure 11-55. — Partial list of parameter values for the Mississippi Delta,

management practice 2.

355

Card 25: The grass strip is not tilled, but the field is tilled above 20
ft from the end of the row. At this crop stage, the critical shear stress of
the tilled soil is estimated to be 0.3 lb/ft2 after subsoiling. The critical
shear stress in the unfilled grass strip is set to 0.7 lb/ft2. The card is
0.0, 0.7, 20.0, and 0.3.

Card 26: Cover stability is assumed; therefore, TCV = 100 lb/ft2.

Cards 27 to 28: The depth to the nonerodible layer is assigned 0.33 ft in
the grass and tilled areas.

Card 29: The flow width in the grass is the entire row width, 3 ft, but
it is 0.5 ft (20 ft up the row) in the row middles that remain after harvest.

Cards 23 to 29: Cards for the field ditch are the same as for example
MP1.

Card 18: 74047, 74078. The cover is developed further in this crop
stage. The field is disked on 078.

Card 19: NC = 1, NP and NM = 0. The soil loss ratio has decreased, but
the contouring and Manning's n factors did not change.

Card 20: SLR = 0.15 for CIN.

Cards 21-22: These cards are absent since the parameters on them did not
change.

Card 23: The only change is in the cover and consolidation of the soil.
Therefore, only Manning's n and critical shear stress change. NN and NCR = 2.
All other N's = 0.

Card 24: 0.0, 0.15, 20.0, 0.10. The Manning's n for the grass is not
changed, but both n's must be reread since Manning's n changes for the tilled
part of the channel. The n on the tilled part increases because resistance to
the flow is assumed to increase from cover in the row middle.

Card 25: 0.0, 0.70, 20.0, 0.4. The critical shear stress for the tilled
portion increases because of consolidation.

Cards 26 to 29: These cards do not exist because their parameters did not
change. Since tillage did not occur, TDN and TDS are not reset. They are not
reset until the next tillage.

On 74079, the field is tilled. This changes the soil loss ratio, Mann-
ing's n, critical shear stress, and depth to the nonerodible layer. The cards
for 74079 to 74107 are for these changes.

Interpretation of Results

The results of three management practices for the Delta site are discussed
in chapter 2.

PLANT NUTRIENTS

Two methods of nitrogen uptake by plants were given in chapter 3 and in

356

volume I, chapter 4. Method 1 is applied on the western Tennessee management
practice 1, and method 2 is applied on the Georgia Piedmont management practice

Nitrogen Uptake Method 1

Western Tennessee Management Practice 1

It was stated in the "Description of Application Sites" section for west-
ern Tennessee that little information is available for the location. It was
stated in chapter 3 that where soil-test data are not available, general infor-
mation from soil surveys may be used in estimating parameter values. Research
data are sometimes available on similar soils and sites. A combination of
sources provided information for estimating parameter values for this applica-
tion.

The following parameter values apply for the nutrient component with
nitrogen uptake method 1.

Cards 1 to 3: WESTERN TENNESSEE MANAGEMENT PRACTICE 1
CONTINUOUS CORN, CONVENTIONAL TILLAGE
NITROGEN UPTAKE METHOD 1 (HYDONE PASSFILE)

Card 4: BDATE = 74001 The beginning date for simulation is January 1,
1974.

FLGOUT = 0 Code for type of output is user specified (coded
for annual summary only).

FLGIN = 0 Code indicates input from hydrology pass file is
in English units.

FLGPST = 0 Pesticides not included in this application.

FLGNUT = 1 Plant nutrients will be simulated.

Card 5: SOLPOR = 0.47 Porosity data for the specific site are unavail-
able. The soil survey data sheets give 1.4 for the bulk density for Loring
silt loam. Porosity = 1 - (BD/2.65) = 0.47 cm3/cm3.

FC = 0.38 Field capacity, volumetric soil water content in
cnr/cm3, was estimated from soil survey data sheets and personal communica-
tion with M. J. M. Romkens, USDA-SEA-AR, Oxford, Miss., who has conducted
research on Loring soils in western Tennessee.

0M = 1.0 Percentage organic matter in the soil survey data
sheets is for surface soil or plow layer. The organic matter content as used
in the nutrient model is for calculations of denitrification in the root zone.
Since data for the profile are unavailable, a reasonable value can be estimated
as half the content in the surface soil. Since the surface value of 2.0% was
obtained, half is 1.0%.

Card 7: OPT = 1 Nitrogen-uptake method 1 is used for this appli-
cation.

357

Card 8: SOLN = 0.24 Initial soluble nitrogen in the top 1 cm of soil
is generally unavailable unless soil tests have been made. For lack of data,
an estimated 5 ppm of nitrogen in the soil water at saturation is reasonable.
A porosity of 0.47 results in 0.235 kg/ha for soluble nitrogen.

SOLP = 0.09 Initial soluble phosphorus in the top centimeter
of soil at saturation would not exceed 2 ppm. With a porosity of 0.47, 2 ppm
would be 2 x 10"6 kg phosphorus/kg water, and 4.7 x 10^ kg/ha water at satura-
tion, or 9.4 x 10"2 kg/ha soluble phosphorus.

N03 = 20.0 Data are unavailable for nitrate in the root
zone. Nitrate in soils at several locations is approximately 20 kg/ha. As
indicated in chapter 4, the default value for N03 is 20, which is used here.

S0ILN = 0.0007 Soil survey data sheets for western Tennessee
give approximately 0.07% nitrogen content in soil, which is in the range of
0.05% to 0.3% for nitrogen in the soil as given in volume I, chapter 4. This
gives 0.0007 kg of nitrogen/kg of soil.

S0ILP = 0.00035 The Loring soils in western Tennessee contain
about 0.03% phosphorus. Conventionally, soil phosphorus is estimated as half
of the soil nitrogen, resulting in a value of 0.00035 kg/kg for S0ILP.

EXKN = 0.07 Extraction coefficient for nitrogen is an empir-
ical coefficient relating nitrogen in the runoff to soluble nitrogen (SOLN) in
the top centimeter of soil. Observations at several research locations have
shown that the value of the coefficient should be in the range of 0.05 to 0.10.

EXKP = 0.07 Extraction coefficient for phosphorus is an
empirical coefficient relating phosphorus in the runoff to soluble phosphorus
(SOLP) in the top centimeter of soil. The coefficient for phosphorus should be
about the same as for nitrogen.

AN = 7.4 The nitrogen enrichment coefficients for sedi-

ment were related to sediment transport in CREAMS, volume III, chapter 12.
Since data are unavailable for Loring soils, the default value of 7.4 is used.

BN = -0.2 The nitrogen enrichment exponents for sediment
were estimated in Vol. Ill, Chap. 12, also. Since data are not available for
Loring soils, the default value, -0.2, is used here.

AP = 7.4 As for nitrogen, the phosphorus enrichment coef-

ficient default value, 7.4, is used.

Card 9: BP = -0.2 The default value, -0.2, is used for phosphorus

enrichment exponent.

RCN = 0.8 Nitrogen concentration in rainfall is read from

the map in figure 1-18 (CREAMS, vol. I, ch. 4).

Card 10: PDATE = 74001 PDATE is the Julian date on which the following
parameters are valid, in this case, the beginning date of simulation.

CDATE = 74309 CDATE is the Julian date on which the model will
stop using the following parameters. This is the day of harvest since potential
mineralization (P0TM) is reset due to increase of organic matter from decaying

358

roots and stover.

Card 15: NF = 1 The number of fertilizer applications made until

CDATE is reached.

DEMERG = 132 The Julian date of plant emergence is assumed to
be 10 days after planting.

DHRVST = 309 Harvest is assumed at the end of October and is
equivalent to the date when LAI goes to zero in the hydrology model.

Card 16: RZMAX = 660.0 Maximum rooting depth is estimated as 660 mm
based on the depth of claypan (personal communication with M. J. M. Romkens,
USDASEA-AR, Oxford, Miss.).

YP ■ 5000.0 The potential yield of corn grain under "ideal"
conditions is about 9,400 kg/ha (CREAMS, vol. I, ch. 4, table 1-11). Informa-
tion from SCS indicates that 80 bu/acre (5,000 kg/ha) is a reasonable potential
yield for the soils and slopes in this field.

DMY = 2.5 Dry matter ratio is the ratio of total dry matter
yield to grain yield. The potential yield of grain and stover is 19480 kg/ha
(CREAMS, vol. I, table 1-11). Roots are considered 20% of the total dry matter
production, giving a total potential of 24,350 kg/ha. DMY then is 24350/9400,
or approximately 2.5.

P0TM = 70.0 Potentially mineralizable nitrogen is calculated
from the organic matter content in the root zone (CREAMS, vol. Ill, Ch. 13).

AWL) = 299.0 Actual water use is the accumulated plant evapor-
ation, in millimeters, for the growing season calculated in the hydrology com-
ponent.

PWU = 299.0 Growing season potential plant evaporation, in
millimeters, is calculated in the hydrology model. The value was calculated by
HYD0NE.

Card 17: CI = 0.0209
C2 = -0.157
C3 = 0.0128
C4 = -0.415

The cubic coefficients and exponents for nitrogen uptake are given by crop
(CREAMS, vol. Ill, Ch. 13, table 3).

Card 18: DF = 74122 The Julian date of fertilizer application is

given in table 11-48 (the same date is used each year).

Card 19: FN = 140.0 The amount of nitrogen fertilizer applied is 140

kg/ha (table 11-48).

FP = 20.0 The amount of phosphorus fertilizer applied is 20
kg/ha (table 11-48).

359

FA = 0.1 Fertilizer was incorporated by disking to a depth
of 10 cm; therefore, the fraction of application in the surface centimeter is
1/10 or 0.1.

When day 74309 is reached in the simulation, card 10 is read for the new
dates of applicability.

Card 10: PDATE = 74310 The Julian date when the new parameters are valid.

CDATE = 75309 The Julian date when the simulation ends with the
following parameters; set at the harvest date in 1975.

Parameters on cards 15 through 19 are updateable to enable the user to
specify actual and potential water use for different crops and different times,
rates, and methods of fertilizer application. The nonupdateable parameters,
such as S0LN, S0ILN, N03, and so forth, are updated automatically by accounting
procedures in the computer program. Updating by the user is unnecessary.
Multiple application of fertilizer during a year can be specified by repeating
cards 18 and 19, as will be shown for the Georgia Piedmont location.
Since updateable parameters in the western Tennessee application are repetitive
for successive years, further discussion is unnecessary. The parameter values
indicate a complete input file for the 3-yr application. Cards 15 through 19
for 1975 follow.

Card

15:

NF =

DEMERG =

132

DHRVST =

309

Card

16:

RZMAX =

660.0

YP =

5000.0

DMY =

2.5

P0TM =

70.0

AWU =

283.0

PWU =

299.0

Card

17:

CI =

0.0209

C2 =

-0.157

C3 =

0.0128

C4 =

-0.415

Card

18:

DF =

75122

Card

19:

FN =

140.0

FP =

20.0

FA =

0.1

The following list shows cards 10 and 15 through 19 for calendar year
1976.

Card 10: PDATE = 75310
CDATE = 76366

Card 15: NF = 1

360

DEMERG =

132

DHRVST =

309

Card

16:

RZMAX =

660.0

YP =

5000.0

DMY =

2.5

POTM =

70.0

AWU =

215.0

PWU =

299.0

Card

17:

CI =

0.0209

C2 =

-0.157

C3 =

0.0128

C4 =

-0.415

Card

18:

DF =

76122

Card

19:

FN =

140.0

FP =

20.0

FA =

0.1

Simulation is to terminate on the last day of 1976. A blank card is in-
serted following card 19. The blank card is read as a zero for CDATE, and sim-
ulation ceases.

Interpretation of Results

Results are summarized in table 11-51 for the 3-yr simulation of plant
nutrients for western Tennessee. This summary includes the water budget, sedi-
ment yield, and plant nutrient budget for the 28-ha area.

Nitrogen additions through fertilization, rainfall nitrogen, and nitrogen
mineralization average approximately 185 kg/ha. Nitrogen uptake was relatively
low for the simulation period due to the low estimated yield of 5,000 kg/ha.
This yield may be low for the climate during the 3-yr period, but it is a real-
istic estimate for the conventional system on steep, eroded soil.

Denitrifi cation is relatively high and probably is greater than expected.
Since immobilization is not considered in the model process, mineralization and
denitrification are higher than expected. The ratio of actual potential plant
evaporation indicates a high soil -water content whereby denitrification is
expected to be high. Nitrate leached, however, is not as high as expected for
the large amounts of annual percolation and denitrification.

Annual summaries may be misleading for mass of pollutants. The sediment
yield for 1974 (table 11-51) was concentrated in two major storm events. Ap-
proximately 70% of the total annual soil loss resulted from these events. The
first storm was the largest and occurred on January 10, which was well before
application of fertilizer. Associated nutrient losses in runoff and sediment
were low. The second largest storm occurred 13 days after fertilization, but
runoff and soil loss were only half that of the January 10 storm. If the mag-
nitudes of the two storms had been reversed, nutrient losses for the year would

361

Table 11-51. — Annual summaries of erosion and water nutrient budgets for west-
ern Tennessee management practice 1

1974

Rainfall (mm) 1 , 79676"

Runoff (mm) 613.9

Percolation (mm) - -218.8

Actual evapotranspi ration (mm) - - - - 907.3
Potential plant evaporation (mm) - - - 299.8
Actual plant evaporation (mm)- - - - - 299.8

Sediment yield (kg/ha) 51,446.

Nitrogen fertilizer (kg/ha)- ----- 140.0

Nitrogen in rainfall (kg/ha) ----- 14.35

Nitrogen mineralization (kg/ha)- - - - 42.85

Nitrogen in runoff (kg/ha) ------- 8.89

Nitrogen in sediment (kg/ha) ------ .75

Nitrogen uptake (kg/ha) 160.26

Nitrate leached (kg/ha) 4.92

Denitrification (kg/ha)- ------- 64.48

Phosphorus fertilizer (kg/ha)- - - - - 20.

Phosphorus in runoff (kg/ha) ------ .54

Phosphorus in sediment (kg/ha) -----' .37

1975

1976

Annual
Average

1,445.3

877.1

1,372.8

357.8

172.0

381.2

195.1

125.4

195.8

895.1

687.5

830.0

299.8

283.2

214.6

265.9

17,709.

6,501.

25,210.

140.0

11.56

7.02

10.98

39.85

34.22

38.97

4.14

2.45

5.16

.23

.12

.37

151.51

137.68

149.82

4.16

5.53

4.87

49.96

32.85

49.10

20.

.31

.15

.33

.11

.06

.18

have been significantly higher. Long-term simulation and analysis of the fre-
quency of occurrence for selected periods of the year therefore are needed.

Nitrogen Uptake Method 2

Georgia Piedmont Management Practice 1

For the Georgia Piedmont example, much data by Smith and others (4) were
used to select parameter values. Indications are given as to how values would
have been assigned without specific published data. The following parameters
are required for the nutrient model when nitrogen uptake is estimated by using
method 2.

Cards 1 to 3: GEORGIA PIEDMONT MANAGEMENT PRACTICE 1
CONTINUOUS CORN, CONVENTIONAL TILLAGE
NITROGEN UPTAKE METHOD 2 (HYDTW0 PASSFILE)

Card 4: BDATE = 74001 The beginning date for simulation is January 1,
1974.

summary).

FLG0UT = 0 Code for type of output desired (coded for annual

FLGIN = 0 Code indicates input from hydrology pass file is
362

nglish units.

FLGPST =

FLGNUT =

Card 5;
(4).

: SOLPOR

= 0

.41

Pesticides not included in this application.
Code to indicate plant nutrient will be simulated.
Soil porosity calculated from a bulk density of

FC = 0.32 Field capacity (4).

0M = 0.65 Percentage of soil organic matter content in
root zone. Half of surface value is used as an estimate.

Card 7: OPT = 2 Nitrogen uptake method 2 is used for this
application.

Card 8: SOLN = 0.20 Initial soluble N in top 1 cm of soil. Gener-
ally unknown unless soil is tested at beginning of simulation. A default value
of 0.2 can be calculated assuming about 5 ppm in the soil solution at satura-
tion.

N03 = 20.0 Initial nitrate in root zone. Default value
used (Ch. 4).

S0ILN = 0.00035 Total nitrogen in root zone. Estimated from
data (4). Units are kilograms of nitrogen per kilogram of soil.

S0ILP = 0.00018 Total phosphorus in kilograms of phosphorus per
kilogram of soil, often nearly half total nitrogen. Assigned value near those
reported by Smith and others (4).

EXKN = 0.10 Extraction coefficient for soluble N in runoff.
An empirical coefficient based on observations using data of Smith and others

(1).

EXKP = 0.10 Extraction coefficient for soluble P in runoff.
Value assigned by same rationale as for soluble N.

AN = 16.8 Coefficient for computing enrichment of N in
sediment by methods in CREAMS, volume III, chapter 12. This value is computed
using data of Smith and others (4).

BN = -0.16 Exponent for computing enrichment of N in sedi-
ment by methods in CREAMS, volume III, chapter 12. This value is computed
using data of Smith and others (4). Without specific data, a default value of
-0.2 would have been assigned.

AP = 11.2 Coefficient for computing enrichment of P in
sediment by method in CREAMS, volume III, chapter 12. This value is computed
using data of Smith and others (4). A default value of 7.4 would have been
used without specific data.

Card 9: BP = -0.146 Exponent for computing phosphorus enrichment in

363

sediment by method in CREAMS, volume III, chapter 12. Value computed using
data of Smith and others (4).

RCN = 0.8 Nitrogen concentration in rainfall reported by

Smith and others (4). Would have been estimated as 1.3 ppm from information in
CREAMS, volume I, chapter 4, figure 1-18.

Card 10: PDATE = 74001 The Julian date for the beginning of simulation
when the following parameters are valid.

CDATE = 74255 The Julian date denoting when the following
parameters are no longer valid (date of harvest to update POTM).

Card 15: NF = 2 The number of fertilizations during the year.

DEMERG = 132 Julian day of plant emergence. Assumed to be 10
days after planting (table 11-45).

DHRVST = 255 Julian day of plant harvest, assumed to be on day
leaf area index = 0.

Card 16: RZMAX = 610.0 Depth of potential root zone; taken here as depth
to B2 horizon, in millimeters.

YP = 5700.0 Potential corn yield in kilograms per hectare
under ideal conditions is 9,400 kilograms per hectare (CREAMS, vol. I, ch. 4,
table 1-11). Since conditions in the Georgia Piedmont are not ideal, potential
is estimated as 5,700 kilograms per hectare.

DMY = 2.5 Dry matter yield ratio, the ratio of total dry
matter production to grain production. DMY is about 2.5 for corn (CREAMS, vol.
I, ch. 4, table 11-11).

POTM = 47.0 Potentially mineralizable nitrogen. Estimated by
method in CREAMS, volume III, chapter 13.

DOM = 60.0 Number of days after emergence until half of
nitrogen is taken up (CREAMS, vol. Ill, ch. 13).

SD = 27.0 Standard deviation of DOM. The number of days
between 50% and 84% uptake (CREAMS, vol. Ill, ch. 13).

PU = 250.0 Potential N uptake by the entire plant, kilograms
per hectare. Based on actual uptake computed from data (4). Range recommended
is 150 to 300 kilograms per hectare (CREAMS, vol. Ill, ch. 13).

Card 18: DF = 74122 Date of fertilization. The first fertilization
was Julian day 122 (table 11-45).

Card 19: FN = 28.0 Amount of nitrogen fertilizer applied, kilograms
per hectare (table 11-45).

FP = 28.0 Amount of phosphorus fertilizer applied, kilo-
grams per hectare (table 11-45).

364

FA = 0.1 Application factor for fertilizer. First appli-
cation incorporated to 10 cm; therefore, FA = 0.1 for first application.

Table 11-45 shows that fertilizer was applied twice in 1974; accordingly,
NF = 2 on Card 15. Cards 18 and 19 must be repeated for each fertilization.

Card 18: DF = 74162 The second fertilizer application in 1974 was
Julian day 162 (table 11-45).

Card 19: FN = 112.0 The fertilizer rate was 112 kg/ha (table 11-45).

FP =0.0 Since the second fertilizer application consisted
of ammonium nitrate, phosphorus was not applied.

FA =1.0 The ammonium nitrate was surface applied, and all
fertilizer is in the top centimeter of soil; thus, application factor is 1.0.

The parameter file is not shown for 1975. Only the updateable parameters
are reset for following years.

Interpretation of Results

Georgia Piedmont is included in the application of nutrients to help the
user select parameter values for nitrogen uptake method 2.

PESTICIDES

Mississippi Delta Management Practice 1

A pesticide application scheme was set up using the tillage/planting pro-
gram (table 11-46) for a pesticide program commonly used in the Mississippi
Delta. Application dates were chosen not to coincide with rainfall, that is,
no application during rainfall. Table 11-52 shows pesticides, application
dates, and assignment of parameter value for management practice 1 in 1974.
For 1975 and 1976, only the application dates were changed as required to match
changes in tillage/planting operation and to avoid pesticide application on
rainy days. Hydrologic information was provided in pass files from the option
2 hydrology model .

Assignment of Parameter Values

Fluometuron--Applied at a rate of 1.5 kg/ha as a preplant incorporated herbi-
cide. Incorporation was assumed uniform to a depth of 10 cm (4 in), that is,
DEPINC = 10, EFFINC = 1. Since the herbicide is applied to soil, S0LFRC = 1,
F0LFRC = 0. Since no initial residues were assumed, F0LRES and S0LRES = 0.
Parameters for foliar washoff are not applicable or required. The program was
written, however, so that zeros can be entered for WSHFRC, WSHTHR, and HAFLIF.
Water solubility of fluometuron is 90 ppm (table 11-40). A value of 0.1 was
assumed for EXTRCT for fluometuron and all other pesticides (CREAMS, vol. I,
Ch. 5). Persistence of fluometuron at the soil surface is described by DECAY =

365

CV O O O O IT!
O COO .-i
C\J coo

ooooooocoo ooooooocoo

■— i^-i^^-ir-l^H>-(^H^-l,— I c>ooooooooo
c\jc\jc\jc\jcvc\jc\jc\jc\jc\j

«d- **- «d- "S-

o <_>

■r- Q£

1 I

>Jo o o o o o

-ol

cocoooooooooooeoooco i/iininininirir'inuiir: o

LOLOLnLOLOLi^LnLOLDLO

OOOOOOOOOO

O O O O O O

cvc\jc\jcvooo->cororoco csjCMCsjcsinroroooroco

o|o o o o o> o

oJo c> o o o o

ooooooocoo OOOOOOOOOO oo

-^O CD OOO O OOOOOOOOOO OOOOOOOOOO o

o o OOO o .— <

oooo oooo

OOOOOO OC5CDOOO

CD CD OOO CD

CD O OOO CD OOOOOOOOOO OOOOOOOOOO O

a. a.

CL LU

fO o

E O O

OOO O OOOOOOOOOO OOOOOOOOOO O

■CO

CM CD

LT> r^ CO

.— i ,— <c\jc\jc\jcnjc\jc\jc\jc\jc\j .— <c\jc\jc\jc\jc\jc\jc\jc\j

Ol =3

& I—

O 4-

s_ r J.

366

0.06, a midrange value (CREAMS, vol. Ill, ch. 17, table 1). The value of 2 for
KD is the mean value for this component (CREAMS, vol. Ill, ch. 19, tables 2 or

4).

Trifluralin--This prepl ant-incorporated herbicide is applied at 1.0 kg/ha by
procedures similar to those used for fluometuron; therefore, DEPINC = 10,
EFFINC = 1, SOLFRC = 1, and FOLFRC = 0. Solubility is 1.0 ppm (table 11-40),
DECAY = 0.07, a midrange value (CREAMS, vol. Ill, ch. 17, table 1). A value of
200 was assumed for KD since some sediment transport of trifluralin is thought
to occur (CREAMS, vol. Ill, ch. 16). Information in chapter 19 and other pub-
lished .information indicate, however, that KD for trifluralin would be about
20, making the sediment transport insignificant. Therefore, a value of 200 for
KD is a compromise between somewhat conflicting observations.

MSMA--Applied three times as a directed postemergence spray at a rate of 0.5
kg/ha each. Since this herbicide is incorporated only by subsequent shallow
cultivation, DEPINC = 1 and EFFINC = 1. Recall that 1 cm is the reference sur-
face depth in the model for calculations of concentrations of surface-applied
pesticides. No significant interception by foliage was assumed; therefore,
SOLFRC = 1, and FOLFRC = 0, with remaining parameters pertaining to foliage set
at 0. Since MSMA is very soluble, a large value (100,000 ppm) was assigned
arbitrarily. DECAY = 0.07 and KD = 4000 were based on observations by R. D.
Wauchope, U.S. Delta States Agricultural Research Center, Stoneville, Miss,
(personal communication).

Diuron--0ne application applied as a direct postemergence spray. Parameter
values are similar to those of MSMA, except S0LH20 = 42, DECAY = 0.185 (mid-
range value from CREAMS, vol. Ill, ch. 17, table 1), KD = 15 (mean value from
CREAMS, vol. Ill, ch. 19, table 3).

Methyl Parathon/EPN--App1 ied together as aerial application to cotton foliage
at rates of 0.5 kg/ha each per application for a total of 10 applications.
DEPINC = 1 and EFFINC = 1 as 1 cm always is used as the reference depth for
computing concentrations at the soil surface for pesticides that reach the soil
surface and are not incorporated physically. Fifty percent of the intended
application is assumed lost off -target by drift and volatilization (CREAMS,
vol. Ill, ch. 18). The 50% intercepted by the target is distributed between
the soil and foliage by FOLFRC = 0.4 and SOLFRC = 0.1 during the first five
applications. Complete canopy closure is assumed at this stage so that FOLFRC
= 0.5 and SOLFRC = 0. The washoff threshold, WSHTHR, was set at 0.2 and 0.3 cm
rainfall, respectively, for these two periods. Organophosphates are removed
readily by rainfall, so that WSHFRC for both methyl parathion and EPN was set
at 0.65 (CREAMS, vol. Ill, ch.18). Foliar half-life values, HAFLIF, of 5 days
for EPN and 3 days for methyl parathion were selected from CREAMS, volume III,
chapter 18, table 2. Rates for soil decay, DECAY, 0.14 for both compounds were
estimated from values in chapter 17 and from personal communications with L. L.
McDowell, USDA Sedimentation Laboratory, Oxford, Miss. KD values for both com-
pounds were estimated using solubility-KD relationships (CREAMS, vol. Ill, ch.
19, figure 6) .

Toxaphene--Al though toxaphene was not applied during the period 1974-76, it was
assumed to be present as a residue in the soil at 3 ppm as a result of past
use. Therefore, APRATE = 0 and S0LRES = 3. DECAY = 0.0014 and KD = 4000 were

367

estimated by L. L. McDowell based on research studies on toxaphene in runoff.

A partial listing of the input parameter file is shown in figure 11-56.
The listing was continued to a point where all initial parameters and the first
set of updateable parameters are given.

CARD
NO

CHEMISTRY PARAMETER DATA

74000
0.440

7
74001

74001

TOXAPHENE

0.000

0.4

74042

FLUOMETURON

0
0.3B0
74001
74041

1.000

0.0

74108

10.000
0.0

74148

1.500
90.0

0
74109

0
74109
TRIFLURALIN

1.000 10.000
1.0

0
74149

0
74149
MSMA

0.500
100000.0

0
74154

0.0

74153

000
0.0

74171

PESTICIDES PARAMETERS - MISSISSIPPI DELTA

MANAGEMENT PRACTICE ONE
CONTINUOUS COTTON - CONUENTIONAL TILLAGE

0 1 0
0.S50
7G3GG

1.000 0.000 1.000 0.000 3.000 0.000 0.000
0.1000 0.0014 4000.0

1.000 0.000 1.000 0.000 0.000 0.000 0.000
0.1000 O.0GO0 2.0

1.000
0.1000

0.000
0.0700

1.000
200.0

0.000 0.000 0.000 0.000

1.000
•0.1000

0.000
0.0700

1.000
4000.0

0.000 0.000 0.000 0.000

Figure 11-56. — Partial list of pesticide input file, Mississippi Delta.

368

74154

MSMA

0.500

1.000

0.000

1.000

100000.0

0.0

0.1000

0.0700

4000.0

74172

74181

74172

MSMA

0.500

1.000

0.000

1.000

100000.0

0.0

0.1000

0.0700

4000.0

74132

74197

74182

DIURON

0.200

1.000

0.000

1.000

42.0

0.0

0.1000

0.1850

15.0

74198

74202

•

74193

METHYL PARATHXON

0.500

1.000

0.400

0.100

GO.O

3.0

0.1000

0.1400

10.0

74193

EPN

0.500

1.000

0.400

0.100

0.5

5.0

0.1000

0.1400

200.0

74203

74209

74203

METHYL PARATHION

0.500

1.000

0.400

0.100

GO.O

3.0

0.1000

0.1400

10.0

74203

EPN

0.500

1.000

0.400

0.100

0.5

5.0

0.1000

0.1400

200.0

0.000 0.000

0.000 0.000 0.000

0.000 0.000 0.G50 0.200

0.G50 0.200

Figure 11-56. — Partial list of pesticide input file, Mississippi Delta'

continued.

369

Interpretation of Results

Annual summaries are given in table 11-53 for the pesticide model output
of the Mississippi Delta application. The model output is not discussed in
detail. A few aspects are explained, however, as examples of the information
contained. In terms of the percent of the amount applied, the greatest predic-
ted pesticide loss was 6.46% for MSMA in 1974. These high losses were caused
by an unusually large amount of rainfall shortly after pesticide application.
About half of the total MSMA loss for the year occurred in a single storm of
6.96 cm one day after pesticide was applied; 3.37 cm became runoff and soil
loss was 3,183 kg/ha. The model output for this storm is in table 11-54. Over
98% of the total storm pesticide loss for MSMA was transported by sediment.

Since rainfall, runoff, and soil loss were less in 1975 and 1976, pesti-
cide runoff losses were reduced.

Toxaphene also is transported primarily by sediment (table 11-54). Losses
were significantly higher in 1974 for several reasons. Soil loss in 1974 was
much greater than in following years. Since no additional toxaphene was ap-
plied, the pesticide residue available to enter runoff declined because of sur-
face depletion by runoff and pesticide decomposition. In an actual situation,
the toxaphene residue at the soil surface would be replaced partially during
major tillage operations that bring soil with higher toxaphene concentration to
the soil surface. The initial soil residue could have been updated at the time
of major tillage operations to partially replace the toxaphene residue avail-
able to enter runoff.

Although it is not the purpose of this example to actually compare differ-
ent management practices and their effects on pesticide runoff potential, it is
obvious that practices that limit soil loss will reduce losses of MSMA and tox-
aphene. Reduction of soil loss has much less effect on the other pesticides.
For relatively nonpersistent pesticides, application timing with rainfall /run-
off occurrence is a dominant factor in relation to time of pesticide applica-
tion.

370

c
c Si

03 T5

i— O
03 T-

O tO

I- <U

CO 03 +->
0) J- c
CL+J 0)

ECO)
i- o to
X o

■o J-

•r- C 0)

CJ O +J

•r- -r- 03

tO 03

<1> i~ M-

Q-4-> «4-

,- C °

E OJ C

3 U 3

E C J-

•r- O

X U C

• I—

-M

S-

CM I

• I

•-■I

03 to

3 4-

c c

r~^ o r*» o

CO CO CTt **

CM
VO

CTi CO

COIOOC^
t£> l£> LO «3"

«tf- CO

"* o

<£> CM *3" CM
CO O CO <£>

cr>

o o

0~> CM

o lo vo ^3-

r-H <d-
r-l KO

co to

CM
CM

CO r*->

CO ^ CO Lf)

,_|

(CVOHfO

CO O

r- 1 CTl <£> ^t

CM CM

VO IT) CM r-H CO HM
"* t-4 LO O i-H CO CO
CM ^" O ^O r^. Ol CT.

MOOiS

LO LO

LO CM CO LO

«d" CTi

HfOHH

CTi r-H

i-H r-. CM t— • CO CM

cm cm «3- r» r». t*«.

OHHH CM CO O

cm o o <— i r». o o

o o o o o o o

>a-coroiD r»«. CM CM

COhh w LO i— i o

CO o o o o <-• o

o o o o o o o

LO CM O VO 1^- CM 1— I

O CM r-H ^J- LO LO O

CM O O CO CO O O

o o o o o o o

(O <D "O

3 E .—

C -r- OJ

C "O •!-

«=C 0) >,

r-» d)
cr. E

r-t o

<- S 3 +j

e c

O T-

•t—

J- r-

•1—

-C

3 03

+->

+J J-

■»->

C r— w

-C

LO Ol 3

C r— 03

o >, i-

r-» E <—

O >> S-

i- -C 03

CT> O M-

J_ J= «3

3 4-> Q.Z

r-l 3 •<-

3+J Q.

r- o;

f— J-

•r- O)

Q S

1—

U_ 1—

s: q s:

c c

O -f-

t. r—

•1—

3 03

tO 4-> J-

+->

1^- <D 3

C r— 03

-C

cr. E .—

o >> s-

^-i o V-

S- .C 03

3 -i-

3 +J Q.2:

p— 5-

c/l

•r- O) Q.

u_ i— 2: q s: uj

h—

371

, —

C1J

**»

■o

s»

+->

-M

(13

i —

-i—

■a

i —

■P

1 —

p —

<+-

M-

S-

i—

•|—

c t— i
•.- >*

-o E
•i- S-
a o

CD t_
Q. O

•r- 03
+J >

O CC OC CO

a < < < o.

ecu »—.

O 4->

•r- E

+j oj

TO E

S- -r-

4-> "O

C 0)

CD oO
U

C C

•r- t_
CD
00 4->
00 "3
03 5

4-> GJ CD
03 i— 13
t- .Q "O
4-> TO -r-
Cr- W
CD •.- CD
C_> TO S-
C >
O 03
CO

c c:

^—,o

•r- CD

03 O

«tf

JC o

«*

•sf

00 -r-

\ o

r»«.

O0 TJ

oi •

•

03 CD

«— o

ttf

*~*l°

i — i

TOO

r~-

^ O

i — i

"crjo

«tf-

1 —

i — I

JO CD CO CO

o ^j- 00 «3"

O "*• i-H «d-

O O CO «3-

O CO CM O

to|o O Is-. o

O CO CM LO

O O LO --H

O t-i «tf- «*

o o <-< o

o o o o

CD O O O O

-a

i —

c_>

+->

-»->

i —

Ll_

1—

t—

372

REFERENCES

(1) Holtan, H. N., C. B. England, G. P. Lawless, and G. A. Schumaker.

1968. Moisture-tension data for selected soils on experimental water-
sheds. U.S. Department of Agriculture, Agricultural Research Service,
ARS 41-144, 609 pp. (Series discontinued; Agricultural Research Ser-
vice is now Science and Education Administration-Agricultural Re-
search.)

(2) Lund, Z. F., and L. L. Loftin.

• 1960. Physical characteristics of some representative Louisiana soils.
U.S. Department of Agriculture, Agricultural Research Service, ARS
4133, 83 pp. (Series discontinued; Agricultural Research Service is
now Science and Education Administration-Agricultural Research.)

(3) Ritchie, J. T.

1972. A model for predicting evaporation from a row crop with complete
cover. Water Resources Research 8(5) : 1204-1213.

(4) Smith, C. N., R. A. Leonard, G. W. Langdale, and G. W. Bailey.

1978. Transport of agricultural chemicals from small upland Piedmont
watersheds. EPA-600/3-78-056. Environmental Protection Agency, Of-
fice of Research and Development, U.S. Government Printing Office,
Washington, D.C. 364 pp.

(5) U.S. Department of Agriculture, Soil Conservation Service.

1972. National Engineering Handbook. Section 4, Hydrology. U.S. Gov-
ernment Printing Office, Washington, D.C. 548 pp.

(6) U.S. Department of Commerce.

1968. Climatic Atlas of the United States. U.S. Government Printing
Office, Washington, D.C.

373

CREAMS

A Field Scale Model for

Chemicals, Runoff, and Erosion From

Agricultural Management Systems

VOLUME III. SUPPORTING DOCUMENTATION

CONTENTS

Chapter Page

1 Time distribution of clock hour rainfall- ------------ 379

--A. T. Hjelmfelt

2 Erosivity "R" for individual design storms- ----------- 386

— K. R. Cooley

3 Estimating Soil Conservation Service runoff curve numbers - - - - 398

on native grazing lands

— C. L. Hanson, E. L. Neff, and A. D. Nicks

4 Residue and tillage effects on SCS runoff curve numbers ----- 405

--W. J. Rawls, C. A. Onstad, and H. H. Richardson

5 Selecting a formula to estimate sediment transport capacity - - - 426

in nonvegetated channels
--C. V. Alonso

6 Contour farming affects runoff patterns and soil movement - - - - 440

--L. D. Meyer

7 Adding erosion from snowmelt to an erosion prediction equation- - 444

— L. D. Meyer

• 8 Modeling erosion and sediment yield from flatland watersheds- - - 446
--C. K. Mutchler and C. E. Murphree

9 Gully erosion -------------------------- 455

--R. F. Piest and E. H. Grissinger

10 Sediment transport capacity of overland flow- ---------- 463

— W. H. Neibling and G. R. Foster

11 Concentrated flow relationships ----------------- 474

--L. J. Lane and G. R. Foster

12 Enrichment ratios for water quality modeling- ---------- 486

— R. G. Menzel

13 Nitrate production, uptake, and leaching- ------------ 493

— S. J. Smith, D. E. Kissel, and J. R. Williams

14 Estimating soluble (PO.-P) and labile phosphorus in runoff 509

from croplands

--L. L. Mcdowell, J. D. Schreiber, and H. B. Pionke

15 Soluble N and P concentrations in surface runoff water- ----- 534

— D. R. Timmons and R. F. Holt

377

CONTENTS

Chapter Page

16 Pesticide concentrations in agricultural runoff: available - - - 544

data and an approximation formula
--R. D. Wauchope and R. A. Leonard

17 Dissipation of pesticides from soils--------------- 560

— R. G. Nash

18 The interception of applied pesticides by foliage and their - - - 595

persistence and washoff potential

— G. H. Willis, W. F. Spencer, and L. L. McDowell

19 Method for distributing pesticide loss in field runoff between - - 607

the solution and adsorbed phase
--H. B. Pionke and R. J. DeAngelis

378

Chapter 1. TIME DISTRIBUTION OF CLOCK HOUR RAINFALL
A. T.'Hjelmfelt-/

INTRODUCTION

Simulation of runoff from rainfall events requires rainfall in time incre-
ments not greater than the time to equilibrium for the watershed. For small
watersheds the time to equilibrium will be measured in minutes, whereas the
easily available rainfall will be available at clock-hour intervals. Time dis-
tribution must be estimated within the clock hour.

CLOCK HOUR AND 60-MINUTE RAINFALL

A rainfall event indicated as occurring in a single clock hour will have
an actual duration of not more than 60 min. Hershfield (2J indicates that the
60-min equivalent is 1.13 times the clock hour rainfall. This is an average of
a value that varies irregularly and unpredictably. The validity of the 1.13
value is indicated in Hershfield's (2) graph of 2-yr events shown in figure 1.

TIME DISTRIBUTION OF RAINFALL

Design Storms

Design storms of a particular return period are generated from intensity-
duration-frequency studies. Results of these studies do not include a time
distribution. The rainfall increments should be arranged to maximize peak dis-
charge but maintain a reasonable sequence within the storm. To achieve this,
early portions of the storm will be subject to interception and depression
storage losses and to higher infiltration losses.

Williams (8) suggests placing the maximum intensity at a point between
one-third and one-half of the storm duration. The other storm increments are
grouped around this value. Hjelmfelt and Cassidy (4) recommend this procedure,
which Kent (6) formalized. Increments of 30 min. were used to form a 24-hr
storm. Maximum intensity was placed near the midpoint of the storm, and re-
maining increments were placed around this value. An average of the resulting
distributions formed the type II distribution for the major portion of the
United States. Another distribution, type I, was generated by a similar method
for the rest of the United States. Type I and type II distributions are shown
in table 1.

1/ Hydraulic engineer, UJDA-SEA-AR, Columbia, Mo.

379

Table l.--Type I and type II
rainfall distributions

Time
(hr)

Px / P24-

Type I

Type II

2.0

0.035

4.0

.076

6.0

.125

7.0

.156

8.0

.194

8.5

.219

9.0

.254

9.5

.303

9.75

.362

10.0

.515

10.5

.583

11.0

.624

11.5

.654

11.75

12.0

.682

12.5

13.0

.727

13.5

14.0

.767

16.0

.830

20.0

.926

24.0

1.000

0.022
.048
.080

.120

.147
.163

.181
.204
.235
.283
.387
.663
.735
.772
.799
.820
.880
.952
1.000

1/ Ratio of accumulated
rainfall to total.

Source: Kent (6J.

Long Duration

Hershfield (3) studied 300 storms
to determine the average time distri-
bution of rainfall within 6-, 12-,
18-, and 24-hr storms. The average
distribution could be displayed as one
curve as shown in figure 2. The type
II storm of Kent (6) also is shown for
comparison.

In discussing the average curve,
Hershfield {3) indicates, "The obvious
limitation of such a curve is that it
conceals the wide variations in the
time distribution and gives no indica-
tion of the distribution from an indi-
vidual storm. Therefore, it would not
be unreasonable to refashion the curve
by rearranging either the duration or

O 1.0 2.0

2-YEAR CLOCK-HOUR RAINFALL (INCHES)

Figure 1. --Relation between 2-yr 60-
min rainfall and 2-yr clock-hr
rainfall .

^^~

yS /

/
/

1
/

HERSHFIELD--

/ /

• /\

*-^"^ l

/ ^- KENT TYPE H -
i i i

20 40 60 80

PERCENT OF TIME

100

Figure 2. --Time distribution of rain-
fall for storms of long duration
as determined by Hershfield (2)
and type II distribution of Kent
(6).

380

storm magnitude increments, because almost any order is realistic." The Bureau
of Reclamation (1_) places the most intense 6-hr at the beginning of a probable
maximum storm.

Short Events

In an intensive study of thunderstorms, the U.S. Weather Bureau (]_) devel-
oped a series of curves describing the time distribution of rainfall in 1-hr
storms. %J The results are shown in figure 3. In these storms the greatest in-
tensity occurs at the beginning. The mass curve of rainfall is the same as the
depth-duration curve.

The curve for 2.01 to 3.00 in/hr storm is decribed by the relation

1-37 T/TTota1
PTotal ~ 0-37 + T/TTotal

Points computed using equation 1 are shown on figure 3,

(1)

100

PERCENTAGE OF STORM DURATION

Figure. 3--Time distribution of 1-hr
storms of various intensities
(_7). Points indicate values
calculated using equation 1.

TIME DISTRIBUTION OF RAINFALL IN
HEAVY STORMS

Huff (_5) published the result of a
detailed analysis of storm patterns in
Illinois. These storms are less than
12 hr, 12 to 24 hr, and greater than 24
hr. They are divided by the timing of
the occurrence of the most intense
rainfall. Thus, the storms are grouped
by most intense portion in the 1st, 2d,
3d, or 4th quartile. Time distribu-
tions are expressed in probabilities.
The 30% probability can be interpreted
that only 30% of the storms will have
this distribution or one of the more
steep distributions. The results of
Huff (5J are shown in figures 4, 5, 6,
and 7. The median, 50%, line is
probably the most useful.

Storms with the most intense por-
tion in the 1st quartile were commonly
of short duration, whereas storms with
most intense portion in the 4th quar-
tile were commonly storms of duration
greater than 24 hrs. The median dis-
tribution from each quartile storm is
compared with the 1-hr storm distribu-
tions of the National Weather Service

2/ In 1970, the U.S. Weather Bureau became the National Weather Service.

381

100

0 20 40 60 80 100

CUMULATIVE PERCENT OF STORM TIME

Figure 4. --Time distribution of rain-
fall with maximum intensity in
1st quartile (5).

100

0 20 40 60 80 100

CUMULATIVE PERCENT OF STORM TIME

Figure 5. --Time distribution of rain-
fall with maximum intensity in
2nd quartile (5).

382

100

CUMULATIVE PERCENT OF STORM TIME

Figure 6.— Time distribution of rain-
fall with maximum intensity in
3rd quartile (5).

100

CUMULATIVE PERCENT OF STORM TIME

Figure 7.— Time distribution of rain-
with maximum intensity in 4th
quartile (5J.

383

100

a. 60
o

— 1 1 1 —

1.0-2.0

N/HR

>%2^^"

2.0-3.0 IN/HR-^

•y

'yf/f\K D

Hf/j b D

/ £ D

/ / //

/ 6 D

I r// °

/& O

If/ o

6 D

■j

f y^

m jt& d

8,12,18,24 HOUR STORMS -
_1 1 1 ' 1

Figu

20 40 60 80
PERCENTAGE OF STORM DURATION

re 8. --Comparison of median
distributions of Huff (_5)
with distributions of the
Weather Bureau (7) and
Hershfield (3).

and with Hershfield's distribution for
long duration storms in figure 8.

RECOMMENDED PROCEDURE

When determining time distribution
of rainfall from clock hour quantities,
the great variability in events must be
recognized. At this time there seems
little advantage in applying an average
distribution to any storm with values
recorded in 3 clock-hours or more. The
average intensity for each hour is prob-
ably the best estimate.

STORMS OCCURRING IN 1 CLOCK-HOUR

Most storms occurring in 1 clock-
hour will last less than 60 min. Using
the total hour as the duration will un-
derestimate the peak discharge. Multi-
plying the clock hour value by 1.13
should result in the equivalent, on a
return period basis, 60-min catch on the
average.

The equivalent 60-min catch be can used with equation 1 to obtain the time
distribution for peak discharge estimates. Thus,

equiv

1.37 (T/60 min)
0.37 +TT760 min;

(2)

in which T is in min

STORMS OCCURRING IN 2 CLOCK-HOURS

If the storm occurs in 2 clock-hours, a little more information is avail
able. Equation 1 can be rewritten

0^37 P/PTotal
1.37--T7-p-fot-

(3)

TTotal
Let the first-hr catch be p\ and the 1 second-hr catch be p£ . Then

Tl _ (0.37) Pl/(pi + p2) .
TTotal ~ 1.37'V-pYfpY + p2y

The result is the fraction of the total duration represented by the first-
hr catch. An additional assumption is needed at this point. If the ratio is

(4)

384

0.50 or greater, one can set T ]_ equal to 60 min and solve the total time

TTotal > wn"icn total will be 120 min or less. If the ratio is less than 0.5,
this process will yield total durations in excess of 2 hr. Recognize that

Tl Tl (5)

TTotal TTotal
and set J\ equal to 60 min to determine Tjota-|.

REFERENCES

(1) Bureau of Reclamation.

1973. Design of small dams. U.S. Department of Interior.

(2) Hershfield, D. M.

1961. Rainfall frequency atlas of the United States. U.S. Department
of Commerce, Weather Bureau Technical Publication 40. (U.S. Weather
Bureau now National Weather Service)

(3)

1962. Extreme rainfall relationships. Journal of Hydraulics Division,
American Society of Civil Engineers Proceedings 88(HY6) : 73-92.

(4) Hjelmfelt, A. T., and J. J. Cassidy.

1975. Hydrology for engineers and planners. Iowa State Press.

(5) Huff, F. A.

1967. Time distribution of rainfall in heavy storms. Water Resources
Research, Vol. 3, 4th quarter, pp. 1007-1019.

(6) Kent, K. M.

1973. A method for estimating volume and rate of runoff in small water-
sheds. Department of Agriculture, Soil Conservation Service SCS-TP-
149 (revised).

(7) U.S. Weather Bureau.

1947. Thunderstorm rainfall. Hydrometeorological Report 5.

(8) Williams, G. R.

1950. Hydrology. ln_ Hydraulic Engineering (Proceedings of the Fourth
Hydraulics Conference), Iowa Institute of Hydraulic Research, H.
Rouse, (ed. ) John Wiley and Sons, Inc.

385

Chapter 2. EROSIVITY "R" FOR INDIVIDUAL DESIGN STORMS
Keith R. Cooley^

INTRODUCTION

In assessing nonpoint source pollution as outlined in Section 208 of the
Federal Water Pollution Control Act Amendments of 1972, Federal and State agen-
cies need to estimate erosion for individual storms since sediments themselves
are pollutants and carry other chemical pollutants. The pollution hazard of
many chemicals applied to agricultural lands is restricted to a short period
immediately after application because most chemicals deteriorate rather rapid-
ly. The first few runoff-producing storms after chemical applications are,
therefore, much more important for assessing possible pollution damage than are
later, possibly more intense, storms. Annual runoff calculations are almost
meaningless for chemical pollutants.

Maps of the erosivity "R" values normally used in the Universal Soil Loss
Equation (USLE) (11) for annual values of erosion do not apply to individual
storms, and actual storm hyetographs often are unavailable. Specifying some
precipitation frequency, or return period, on which to base estimates frequent-
ly is desirable for designers. The method presented here provides R-values for
individual storm events of any selected standard design frequency and duration
(7, 8) for any of the four types of storms defined by the Soil Conservation
Service (SCS) ( 9J . A general equation relating maximum 30-min intensity for
storms of any duration and volume of total precipitation also is presented for
each type of storm.

Procedure

Although any rainfall distribution could be used, the SCS storm types I,
IA, II, and 1 1 A rainfall distributions (9j were used because they are probably
the most common. By normalizing the time axis of these four rainfall distribu-
tion plots, a table of normalized time vs. normalized rainfall was developed
for each type of storm. The rainfall distribution within any selected frequen-
cy of design storm was determined by multiplying total storm rainfall by the
fractional rainfall increments, corresponding to selected uniform time incre-
ments, for the SCS type storm desired. Intensity in inches per hour (1 in/hr =
0.007 mm/s) was calculated for each increment by dividing the rainfall occur-
ring during that increment by the incremental time value.

Energy per inch (1 in = 25.4 mm) for each rainfall increment was calcula-
ted according to the relationship:

\.l Hydrologist, USDA, SEA-AR, U.S. Water Conservation Laboratory, Phoe-
nix, Ariz.

386

E = 919 + 331 log10 I

(1)

where E = energy in foot tons per acre per inch (1 ft-ton/acre-in = 26.38
J/nr) , and I = incremental rainfall intensity in inches per hour (1 in/ hr =
0.007 mm/s), as calculated previously (_10 ) . The energy per increment was de-
termined as the product of each energy-per-inch value and the corresponding in-
crement of rainfall in inches (1 in = 25.4 mm). The product of the sum of the
individual energy-per-increment values and the maximum 30-min rainfall intensi-
ty, divided by 100, provides the erosivity factor "R" for the type and frequen-
cy of design storm selected. The maximum 30-min intensity is a function of
storm type, and expressed by a general equation. Maximum 30-min intensity in
inches per hour (1 in/hr = 0.007 mm/s) equals

Imax = (P) (aDp)

(2)

where P = total storm rainfall in inches (1 in = 25.4 mm), D = storm duration
in hours, and a and e are constants for any given storm type. The values of a
and 6 are presented in table 1 for each of the four SCS storm types used.

Using a similar approach, Ateshian
(_1) presented a method of determining R
values for individual storms of any
duration, and 24-hr storms, for types I
and II distributions. His main empha-
sis, however, was to develop a rela-
tionship between 2-yr, 6-hr rainfall
and the average annual erosion index R.
His general equation for individual
storms of any duration was:

EI
100

= R

a'P

2.2

(3)

Table 1.— Values of a and 3
for use in equation 2 for
each type of SCS storm

Type of

Storm

Coefficients
a 6

IA
I

II
IIA

1.36 -0.56
1.51 .40
1.68 .25
1.82 .136

where P and D are as defined above and a' and b' are constants depending on the
type of storm.

In this analysis the general equation was found to be:

EI -n-aPf(D>
100 Db

(4)

where P and D are as defined previously and a and b are constants depending on
type of storm. The power to which rainfall P is raised also is a function of
duration f(D). The function f(D) was evaluated by regression analysis using
values from the four storm types and seven storm durations. The best fit rela-
tionship for all storm types had a regression coefficient r2 of 0.98 and was
found to be:

387

f (D) = 2.119D*0086. (5)

Substitution into equation 4 yields:

EI p2.119D-0086

TOO ~ R jjb » (6)

which can be used to determine individual storm R values for storms of any
duration and total precipitation. Maximum values of total rainfall for the
different durations were based on reports from U.S. weather stations (3_, 6).
Table 2 shows the coefficients a and b in equation 6 for each SCS type storm.

Figure 1 shows the percent
difference between R values calculated Table 2. --Values of a and b in
by the Ateshian method (eq. 3) and the equation 5 for each SCS
method presented here (eq. 6) as a type of storm

function of storm duration and total

precipitation for the type II storm

distribution. A positive difference Type of Storm Coefficients
means that the Ateshian method
computes an R value greater than that
computed by equation 6. The percent
difference is greater for short
duration storms of high magnitude and
is less than 10% for 12- and 24-hr
storms of any storm magnitude shown
(fig. 1). Since Ateshian used the
24-hr storm to develop his relation-
ship and rounded the coefficient 2.2 to the nearest tenth, the 24-hr values
should be nearly the same as those produced by equation 6.

.98

0.7488

.03

.5780

.90

.4134

II A

.51

.2811

Renard (£) , in discussing Ateshian' s paper ( 1) , superimposed typical
short-duration, high-intensity air-mass thunderstorm depth-duration curves for
11 storms on a plot of the type I and II distributions presented by Ateshian.
Renard suggested that the type I and II distributions poorly represent these
thunderstorm distributions. He did not normalize the time scales, however,
and only one or two of the longer duration storms appear similar to the type I
and II curves. Using 9 of the 11 storms described by Renard and the 4 SCS
curves, plots normalized in both precipitation and time are presented in fig-
ures 2A and 2B for Walnut Gulch, Ariz., and Alamogordo Creek, N.Mex. respec-
tively. These plots show that the SCS curves more nearly represent the actual
storm distributions when time and precipitation are normalized. Even when the
most intense storms are selected (as shown here), distributions vary widely
and include several storm types. The plots show that the four SCS curves do
not cover all possible distributions. A better measure of their representa-
tiveness to erosivity would be to compare actual storm R values with those
produced by the SCS type storms.

In a separate discussion of Ateshian's paper ( Jj , Renard and Simanton (5J
prepared a table showing the the actual computed R values for the same 11
storms. They compared R values calculated using Ateshian's equations for 24-

388

-20

10 0 10 20 30 40 50
PERCENT DIFFERENCE

Figure 1.— The percent difference in R value
calculated by the Ateshian and the Cooley
methods for SCS type II storms of various
durations (D) as a function of total
storm precipitation (P). A positive dif-
ference indicates that the value calcula-
ted by the Ateshian method was larger.

hr storms, and for storms of any duration, to the actual R values. As
expected, the calculated R values using the equation for 24-hr storms are
considerably in error and different from the values calculated by Ateshian's
other equation (eq. 3), except for the 25.82-hr storm when both produced
essentially the same results since this storm lasted nearly 24 hr.

Using the data of Renard and Simanton ( 5J , table 3 compares actual R val-
ues with those determined for the four types of storms, using equation 6. Ta-
ble 3 also shows the values obtained using Ateshian's method for storms of any
duration, the type of storm most nearly matching actual values, and the per-
cent-error for each. As shown, some types of storms may be predominant in an
areat but most areas exhibit a large range of variability about this type (2).
At Walnut Gulch, and especially Alamogordo Creek (table 3), the type 1 1 A ston
is predominant, but essentially the entire range of types is represented
Using equation 6, all types of storms can be considered in a design procedu'
and which type, or under what set of circumstances, the most critical cone'
tions occur and how often can be determined.

389

0 10 20 30 40 50 60 70 80 90 100
PERCENT DURATION (D)

100

£ 80

-I 70

-I
<
U. 60

< 50

& JO x

9 to /

0 10 20 30 40 50 60 70 60 90 100
PERCENT DURATION (D)

Figure 2. --Plot of normalized rainfall distribu-
tions for actual storm data (dashed lines),
and SCS type IA, I, II, and 1 1 A normalized
storm distributions (solid lines) at (A) Wal-
nut Gulch, Ariz., and (B) Alamogordo Creek,
N .Mex.

390

ocooiai

o a-i -— i lo

LO LO CO CTi CTi CO
CM CM CM CM

mh Org-*

MONICO
^H rH CM ,H (\1

CO CO LO CM CM

OODOHDN CO
CO CM CM CSJ .—I

iDco^otnai

OlDOl'-H'-HD
CO CM CM ^- i— I .— I

CM CM CM CM

LO -^- CO CO LO

CO LO =d" O CM
CM CM CO CO CO

LO CM CO

•a- m n
lo co oo

en en o. cn< —

=3 =5 O) ^ =3

< < W<T

Or — LO LO LO CM
CO LO LO LO CT* r^
CTt O^i CTi CTt i — I CT»

C i— C CTi =>

3 3 =J 3 r_3

■"3 rO <~S <C

i — c:

391

Table 4 is a similar analysis of 31 storms selected to cover a large range
in duration and rainfall from four sites in Hawaii. Type I and IA storms are
dominant, although all four types are encountered with type 1 1 A occurring only
once. Figure 3 shows the contrast in the distribution of occurrence of each
type between the southwestern United States and Hawaii. These data were ob-
tained by combining Walnut Gulch with Alamogordo Creek and by combining the
four Hawaiian sites.

RESULTS AND DISCUSSION

In addition to these simple equations, a computer program incorporating
the relationships allows one to quickly and easily determine the erosivity fac-
tor R for any type and design storm selected (this program is available on re-
quest to the author). The effects of types of storms on the erosion potential
of any given site also can be determined easily. Using the same design storm
of 2-yr frequency and 6-hr duration at Laupahoehoe, Hawaii, for example, the R
value ranged from 192 to 677 as type of storm changed. Table 5 shows the com-
puter program for the type IA storm. This table gives the incremental values
calculated by the preceding method.

Although some areas of the world are subjected to storms of a predominant
type, most areas are subjected to a variety of storms with an annual average
approaching one type but with considerable variation occurring about this mean
( 2J . Using the preceding method, the designer can decide what type of storm to
use and how much the range in R values will be under his conditions. A type I

o
o
10

°5

□ HAWAII

Q SOUTHWESTERN U.S.

i n

STORM TYPE

Figure 3. --Distribution of SCS storm types in
Hawaii and the southwestern U.S.

392

Table 4.— Actual rainfall erosion index for individual storms compared with

values for SCS type IA, I, II, and 1 1 A distributions in Hawaii

:4/

Watershed

and
storm data

Calculated R

Duration Rainfall Actual R

Best
fit

(hr)

Laupahoehoe

(in)

(ft-t/ac)

342/73 1.83

54/72 2.92

304/72 3.33

60/76 4.34

18/74 5.42

32/74 9.67

78/74 12.17

51/73 14.75

323/73 --19.92

315/73 30.25

6/75 75.87

Waialua Pineapple

208/74 1.17

109/74 3.75

277/72 5.33

77/74 11.00

108/74 18.58

Mil ilani

109/74 1.58

184/74 - 1.83

264/74 2.17

262/74— 2.75

32/74 4.00

38/76 - 6.91

38/76- -— 9.69

37/76 13.75

31/75— -—16.57

Kunia

179/77 - 5.37

132/77—- 7.83

337/77 8.31

31/75 19.89

37/76 —20.67

38/76— 30.71

1.50
1.30
1.50
1.91
3.40
3.60
5.40
5.40
7.20
17.60
22.28

1.50
2.10
2.00
3.90
15.30

1.30
2.30
1.80
4.30
2.40
2.50
2.24
2.66
4.76

1.37
.85
2.08
4.11
3.12
5.97

26
23

14
29
64
85

148
52
125
492
234

22
33
34
75
825

16
49
32
234
36
59
23
14
80

7
2
13
30
16
49

79 124
64 112

77 137 245

67 123 228
101 195 380
528 1094 2283
468 1135 2756

27 32 40

24 34 51

16 25 40

41 71 126

546 1042 2007

16 20 26

49 62 82

25 34 46

138 189 266

30 44 66

22 35 58

14 23 40

15 28 50

88 166

58 112

31 61

103 215

II2/
IIA^7
IA

I
I
I

If Calculated by equation 6.

1J Calculated R value for type IIA storm = 8.

Note: 1 in = 25.4 mm; 1 ft-t/ac = 0.67 J/m2.

393

Table 5.— Typical output from computer, program for Laupahoehoe, Hawaii:
2-year, 6-hour type IIA storm-

Energy

Time

Rain

Intensity

per

Increment

acre

Increment

(hr)

(in)

(in/hr)

(ft-t/ac-in)

(ft-t/ac)

0.125

0.019

0.15

644

.125

.019

.15

644

.125

.019

.15

644

.125

.025

.20

686

.125

.025

.20

686

.125

.025

.20

686

.125

.031

.25

718

.125

.044

.35

766

.125

.057

.45

802

.125

.088

.71

866

.125

.372

2.97

1073

399

.125

3.723

29.79

1404

5227

.125

.315

2.52

1049

330

.125

.139

1.11

931

129

.125

.126

1.01

917

116

.125

.101

.81

885

.125

.088

.71

866

.125

.081

.66

855

.125

.075

.60

844

.125

.063

.50

818

.125

.050

.40

785

.125

.050

.40

785

.125

.044

.35

766

.125

.044

.35

766

.125

.044

.35

766

.125

.038

.30

744

.125

.038

.30

744

.125

.038

.30

744

.125

.038

.30

744

.125

.038

.30

744

.125

.038

.30

744

.125

.031

.25

718

.125

.031

.25

718

.125

.031

.25

718

.125

.031

.25

718

.125

.031

.25

718

.125

.025

.20

686

.125

.025

.20

686

.125

.025

.20

686

.125

.025

.20

686

.125

.019

.15

644

.125

.019

.15

644

.125

.019

.15

644

.125 •

.019

.15

644

.125

.019

.15

644

.125

.019

.15

644

.125

.019

.15

644

.125

.013

.10

586

1/ Depth = 6.3 in; erosivity (R) = 677.05; maximum 30-min intensity =
9.10 in/hr.

394

storm during winter, when the soil is nearly bare, may be more critical than a
type II storm occurring in summer when the soil surface is protected by a full
plant canopy.

In areas like Hawaii, where sugarcane and pineapple can be harvested any
time during the year, the same type of analysis may help managers determine the
best harvest schedule to minimize erosion. Fields most susceptible to erosion
because of steepness or soil type, for example, can be harvested during the
least critical period, when storms with high erosivity potential are least
likely to occur. Figure 4 shows the relationship between individual storm EI
or R data and day of the year, using 1974 storm rainfall data from Laupahoehoe,
Hawaii. The best harvest date for susceptible fields, as shown in figure 4,
would be about day 120.

100

i 1 1 1 1 1 1 r

TOTAL El 951
MAX. El 148
RATIO MAX./ TOTAL 0.16

120 150 180 210 240 270 300 330 360
JULIAN DAY

Figure 4.— Relation of individual storm erosivity (EI)
with time of year (expressed as Julian day) for a
study site near Laupahoehoe, Hawaii, in 1974.

395

CONCLUSIONS

The method presented here to determine erosivity R values for individual
storms provides an easy, rapid procedure to assess erosion or pollution poten-
tial for a storm when combined with erosion or chemical transport relation-
ships. This method also allows designers of conservation measures to determine
the range of R values that might be encountered and the most critical combina-
tion of storm type and soil conditions. Managers in some areas also may use
the distribution of erosivity with time to minimize soil losses by proper ad-
justment of harvesting schedules.

Although the four SCS type curves do not encompass all possible storm dis-
tributions, results for two widely different areas (southwestern United States
and Hawaii) indicate that R values calculated using these curves would be ac-
curate enough for most designs. The type 1 1 A distribution produced R values
close to actual values calculated from selected intense thunderstorms in the
southwestern United States.

The author appreciates the assistance of Tom Hansen who wrote the program
for this procedure and Dave Larson who did much of the data processing.

REFERENCES

(1) Ateshian, J. K. H.

1974. Estimation of rainfall erosion index. Journal of the Irriga-
tion and Drainage Division. Proceedings of the American Society of
Civil Engineers 100(IR3) :293-307.

(2) Hershfield, D. M.

1962. Extreme rainfall relationships. Journal of the Hydraulics Di-
vision. Proceedings of the American Society of Civil Engineers
88(HY6):73-92.

(3) Niedringhaus, T. E.

1973. Rainfall intensities in the conterminous United States and
Hawaii. Special Report ETL-SR-74-3, U.S. Army Engineers, Topogra-
phic Laboratories, Fort Belvoir, Va. p. 36.

(4) Renard, K. G.

1975. Discussion of estimation of rainfall erosion index. Journal of
the Irrigation and Drainage Division. Proceedings of the American
Society of Civil Engineers 101( IR3) :240-241 .

(5) , and J. R. Simanton.

1975. Discussion of estimation of rainfall erosion index. Journal of
the Irrigation and Drainage Division. Proceedings of the American
Society of Civil Engineers 101 ( IR3) :240-241 .

(6) Riordan, P.

1970. Weather extremes around the world. Technical Report 70-45-ES,
U.S. Army Natick Laboratories, Natick, Mass. 38 pp.

396

(7) U. S. Department of Agriculture, Soil Conservation Service.

1970. Estimating peak discharges for watershed evaluation storms and
preliminary designs. Technical Service Center Technical Note - Hy-
drology-P0-2, 6 pp.

(8) U. S. Department of Commerce, Weather Bureau.

1961. Rainfall-frequency atlas of the United States. Technical Paper
No. 40, 115 pp.

(9)

1962. Rainfall -frequency atlas of the Hawaiian Islands. Technical Pa-
per No. 43, 60 pp.

(10) Wischmeier, W. H., and D. D. Smith.

1958. Rainfall energy and its relationship to soil loss. Transactions
of the American Geophysical Union 39(3) : 285-291*

(11) , and D. D. Smith.

1978. Predicting rainfall erosion losses--a guide to conservation
planning. U. S. Department of Agriculture, Agriculture Handbook No.
537, 58 pp. (Purdue Agricultural Experiment Station cooperating.)

397

Chapter 3. ESTIMATING SCS RUNOFF CURVE NUMBERS ON NATIVE GRAZTNG LANDS
C. L. Hanson, E. L. Neff, and A. D. Nicks^

INTRODUCTION

The United States contains 517 million acres of privately controlled lands
classified as native grazing land (rangeland, grazable woodland, and native
pasture) (8_) . This area constitutes 36% of the land area in the United States.
Approximately 64% of the total grazing land is on land classified as having an
erosion hazard. Conservation treatment and improvement of existing cover are
needed on 71% of these lands; brush control and reestabl ishment of cover are
needed on another 10%, and 5% of this land is not suitable for treatment (_4 ) .

Water-pollution hazard from native grazing lands is limited mainly to
transport of sediments eroded from them. Chemical applications to these graz-
ing lands consist primarily of fertilizer to reestablish cover and herbicides
and pesticides to control brush and pests in some areas. Using mathematical
models to estimate the best management practices is important, however, because
of the large area in grassland that is susceptible to erosion. Estimating the
runoff potential of grazing land sites also may be difficult because of the
varied soil, cover, and grazing intensity. This chapter gives some methods and
data that the planner can use in estimating runoff potential by the Soil Con-
servation Service (SCS) runoff curve number method. Specifically, the data and
procedures are given as a guide to estimate the averaqe curve number parameter
required by Hydrology Option One of the CREAMS model.

PROCEDURES OF THE SOIL CONSERVATION SERVICE

The SCS curve number method was used to estimate direct runoff in the re-
port "Control of Water Pollution from Cropland" (3_) . The procedures used were
developed to determine the curve number (CN) for many agricultural practices
(5). The following summary of procedures was used by State and Federal agen-
cTes to estimate curve numbers for rangeland. Suggested curve numbers for the
Northern Great Plains also are given.

The SCS (5_) and the Bureau of Land Management (_U) have graphs to esti-
mate curve numbers for Pinyon-Juniper and sagebrush cover classes. The Bureau
of Land Management (11) also has a graph for grassland. These graphs are based
on the hydrologic soil groups and percentage of cover. The percentage of cover

]J Agricultural engineer, USDA-SEA-AR, Northwest Watershed Research Cen-
ter, Boise, Idaho; hydraulic engineer, USDA-SEA-AR, Northern Plains Soil and
Water Research Center, Sidney, Mont.; and agricultural engineer, USDA-SEA-AR,
Southern Great Plains Watershed Research Center, Chickasha, Okla.

398

classifies the cover in poor, fair, and good hydrologic condition. SCS gives
two procedures for determining the hydrologic cover condition of rangeland (_5_,
ch. 8). The SCS also developed "Hydrology Technical Note PO-7," a photographic
catalog illustrating range sites and hydrologic conditions (6.)*

The SCS in Arizona and New Mexico developed a figure representing curve
numbers for their hydrologic conditions. This figure, based on figures 9.5 and
9.6 in the SCS's National Engineering Handbook, section 4 (5J , expresses the
runoff curve numbers as a function of cover density and hydrologic soil type
for various vegetation types [29 1_ ) . The SCS in Arizona also developed a meth-
od of adjusting curve numbers for storm duration {ls JL?_ ) • The SCS in Wyoming
developed a table of soil cover complex numbers derived from range sites and
condition of cover (10) .

Table 1 lists runoff curve numbers in relation to range sites and condi-
tion of cover. These data were adapted from the Wyoming SCS table for use in
the Northern Great Plains. The values in table 1 were verified from SEA-AR wa-
tershed data in western South Dakota, southeastern Montana, and northeastern
Wyoming.. They represent antecedent moisture condition I. The range condition
classes of fair-good, high-fair, and excellent in the Wyoming SCS table were
changed to poor, fair, and good to coincide with tables 8.1 and 8.2 of the Na-
tional Engineering Handbook (_5) .

Using a climate index (9J the SCS in Texas developed a procedure to adjust
curve numbers. This procedure is for all types of soil-cover complexes and al-
so includes grazing land conditions that are prevalent in large areas of west-
ern Texas. Figure 1 shows the adjustment to be made to the curve number of a
given soil -cover complex. To use this procedure, select the condition I and
condition II curve number from the National Engineering Handbook, section 4
(_5) . Then select the appropriate isogram values from figure 1 for the range
site. Calculate the average curve number, using the isogram and the curve num-
bers from the SCS table. An example of this calculation is:

CNave = CNI + 0.20 (CNII-CNI)

where CNI and CNII are curve numbers for antecedent moisture condition I and II
for the soil -cover complex, and CNave is the average adjusted curve number for
the site.

This procedure is recommended to obtain the average runoff condition curve
number for emergency spillway and freeboard hydrographs. No runoff curve num-
ber less than 60 is used as a result of this adjustment unless the unadjusted
soil-cover complex number is less than 60. When this occurs, no downward ad-
justment is used.

USDA-SEA-AR WATERSHED CURVE NUMBER

Data from research watershed range sites in the northern and southern
Great Plains were used to calculate representative values of average curve num-
bers. These values are shown for the model user's reference in selecting the
average curve number for Hydrology Option One. Values for the northern Plains,

399

Table 1. --Runoff curve numbers derived from range sites and condition of cover
for antecedent moisture condition I.

Range condition

Range Site
Poor Fair Good

Wetland 95

Very shallow 95

Sal ine subirrigated 90

Sub irrigated 90

Shale— 90

Dense clay 90

Alkali clay 90

Sal ine upl and 90

Igneous 90

Shallow clayey 85

Shallow sandy 80

Shallow loamy 80

Shallow igneous 80

Steep clayey 80

Clayey 80

Gravelly loamy 80

Steep loamy 80

Overflow 80

Loamy overflow 80

CI ayey overflow 80

Coarse upland 80

Limy upl and 80

Shallow breaks 80

Stony 80

Steep stony 80

Lowland 80

Sal ine lowland 80

Loamy lowl and 80

Loamy 80

Sandy lowl and 75

Sandy 75

Gravelly 70

Sands 70

Choppy sands 70

Note: As sites conditions are general, the curve number should be adjust-
ed (interpolated) for each particular site based upon a field investigation.

given in table 2, compare closely with estimates given in table 1. Values for
the southern Plains are given in table 3, which compares two types of rangeland
conditions that are prevalent in this area. These conditions are native range-
land never cultivated and native rangeland formerly cultivated, abandoned, and
left to revert back to native grass. Three watersheds in the table 3 represent

A 00

Table 2.— SCS curve number from selected USDA-SEA-AR watersheds in the northern Great

Plains

Watershed

Watershed .
number Area

Range type C(

Range

Hydro logic

SCS

curve n

umber

locat

ion

)ndition

group

Low

Average

High

(acres)

Hasting

, Nebr.

1H 3.62

Native meadow

Fair

Hasting

, Nebr.

2H 3.40

Native meadow

Fair

Hasting

, Nebr.

18H 3.74

Native pasture
(heavy grazing) ,

Fair

Ekalaka

, Mont.

1 2.00

Saline-upland
range site.

Poor

Ekalaka

, Mont.

2 2.00

Panspots

Poor

Ekalaka

, Mont.

3 2.00

Panspots

Poor

Cottonwood, S.

Dak. 4H 8.57

Pierre shale

Fair

(heavy grazing).

Cottonwood, S.

Dak. 5M 8.57

Pierre shale

Fair

(medium grazing]

Cottonwood, S.

Dak. 6L 8.99

Pierre shale

Good

(light grazing).

Newell,

S. Dak.

2—115.00

Medium-textured
soils (mixed
range sites).

Poor

Newel 1,

S. Dak.

55—41.40

Medium-textured
soils (mixed
range sites) .

Fair

Newell ,

S. Dak.

7—160.00

Medium-textured
soils (mixed
range sites).

Poor

Newell ,

S. Dak.

12—90.00

Fine-textured
soils (mixed
range sites) .

Poor

Newell,

S. Dak.

13—60.00

Fine-textured
soils (mixed
range sites).

Poor

Newell,

S. Dak.

14—35.00

Fine-textured
soils (mixed
range sites).

Poor

Newell,

S. Dak.

15—115.00

Fine-textured
soils (mixed
range sites).

Poor

Newell,

S. Dak.

51 7.90

Sandy range
sites.

Fair

Newell,

S. Dak.

53—11.30

Sandy range
sites.

Fair

401

Table 2.--SCS curve number from

selected USDA-SEA-AR watersheds in the northern Great
Plains—Continued

Watershed
location

Watershed .
number Area

«^type co^on

Hydro logic
group

SCS i
Low ,

curve number
Average High

(acres)

Newell, S. Dak.

55—

-16.50

Sandy range
sites.

Fair

Newell, S. Dak.

P5—

—8.00

Panspots

Fair

Newell, S. Dak.

P6- —

-13.20

Panspots

Fair

Newell, S. Dak.

P7— -

—7.25

Panspots

Fair

Newell, S. Dak.

P8- —

—6.42

Panspots

Fair

Newell, S. Dak.

P9— -

—6.96

Panspots

Fair

Aladdin, Wyo.

1 — .

—7.70

Silty range
site.

Fair

Aladdin, Wyo.

2— .

—8.20

Silty range
site.

Fair

Aladdin, Wyo.

3— ■

-11.60

Shallow range
site.

Fair

Aladdin, Wyo.

4— .

—2.50

Shallow range
site.

Fair

Reynolds, Idaho

1— .

-205.00

Summit water-
shed (mixed
range site).

Poor

Reynolds, Idaho

2— .

-33.00

Lower sheep
(mixed range
site).

Poor

Reynolds, Idaho

3— •

-306.00

Murphy (mixed
range site).

Fair

Reynolds, Idaho

4___.

-100.00

East Reynolds

Fair

Mt. (mixed range

site).

Table 3. — SCS

curve numbers from selected USDA-SEA-AR watersheds in

the

southern

Great Plains

Watershed
location

Watershed .
number Area

Retype Jj

Hydrologic
group

SCS curve number
Low Average High

(acres)

Guthrie, Okla.

W-I —

-2.50

Virgin native
grass.

Good

Guthrie, Okla.

W-II-

-5.09

Virgin native
grass.

Good

Guthrie, Okla.

W-III-

-9.09

Formerly culti-
vated; eroded.

Fair

Guthrie, Okla.

W-IV-

-13.40

Formerly culti-
vated; eroded.

Fair

Guthrie, Okla.

W-V—

-15.70

Formerly culti-
vated; eroded.

Fair

Guthrie, Okla.

PL, L-

-5.62

Native woodland

Fair

Guthrie, Okla.

PL, J-

-5.28

Severly eroded

Poor

Guthrie, Okla.

PL, ISA-

-3.13

Formerly culti-
vated; terraced.

Good

Guthrie, Okla.

PL, 13-

-3.21

Gullied; reformed Good

Chickasha, Okla.

R-2—

-24.08

Sandy range site

Fair

Chickasha, Okla.

R-5-

-23.72

Virgin rangeland
site.

Good

Chickasha, Okla.

R-7 —

-19.19

Formerly culti-
vated; treated

Poor

402

106°

104°

102°

TEXAS

100°

§6°

94°

36°

32°
30°
28°
26°

(

"^YVuYVT V r^-X

i l.\l.uAHk

X £A--*^\^ V-^V\_aA.\

l\M

YujSn MI + .4 .

\lkvy (iii-ii:

<kX\/cA^\74. ao m-i)

yr \ i + .60{ii-i)

yvX^V

^-I+.40 (II-I)

M + .20 (II-I)

SCALE IN Ml

.ES

Figure 1. --Adjustments for runoff curve number in Texas (9).

reclaimed eroded, gullied lands (W-III, W-IV, and W-V at Guthrie, Okla.), three
represent virgin native conditions (W-I and W-II at Guthrie, and R-5 at Chick-
asha, Okla.)> and two represent no treatment after abandonment with natural re-
version to native grass.

REFERENCES

(1) Malone, J.M.

1972. Hydrologic design manual for drainage areas under 25 square
miles. [Preliminary draft] U.S. Department of Agriculture, Soil Con-
servation Service.

(2) Simanton, J.R., K.G. Renard, and N.G. Sutter.

1973. Procedures for identifying parameters affecting storm runoff
volumes in a semiarid environment. U.S. Department of Agriculture,
Agricultural Research Service, Western Region, ARS-W-1, 12 p. (Series
discontinued; Agricultural Research Service is now Science and Educa-

403

t ion Administration-Agricultural Research.)

(3) Stewart, B.A., D.A. Woolhiser, W.H. Wischmeier, J.H. Caro, and M.H. Frere

1976. Control of water pollution from cropland, Vol. II - An overview.
U.S. Department of Agriculture, Agricultural Research Service, ARS-H-
5-2, 187 pp. (Series discontinued; Agricultural Research Service is
now Science and Education Administration-Agricultural Research.)

(4) U.S. Department of Agriculture.

1976. National inventory of soil and water conservation needs. U.S.
Department of Agriculture Statistical Bulletin No. 461.

(5) U.S. Department of Agriculture, Soil Conservation Service.

1972. SCS National Engineering Handbook, Section 4, Hydrology. 458
pp.

(6)

1973. Photographic catalog of range sites and their hydrologic condi-
tion. U.S. Department of Agriculture, Soil Conservation Service, Hy-
drology Technical Note PO-7, 23 pp.

(7)

1973. Peak rates of discharge for small watersheds. _In_ Engineering
field Manual for conservation practices, ch. 2. (revised 10/73 for
New Mexico) .

(8)

1976. National Range Handbook Notice-1, Washington, D.C.

(9)

1978. Engineering-Hydrology Memorandum TX-1, Temple, Tex.

(10)

1978. Runoff and yield determination procedures. U.S. Department of
Agriculture, Soil Conservation Service, Technical Note Engineering
18, page 16, Table 3. Casper, Wyo.

(11) U.S. Department of Interior, Bureau of Land Manaqement.

1969. Bureau of Land Management Manual 7313, cover.

(12) Woodward, D.E.

1973. Runoff curve numbers for semiarid range and forest conditions.
American Society of Agricultural Enqineers Paper 73-209, St. Joseph,
Mich.

404

Chapter 4. RESIDUE AND TILLAGE EFFECTS ON SCS RUNOFF CURVE NUMBERS

W. J. Rawls, C. A. Onstad and H. H. Richardson^'

ABSTRACT

The effect of conservation tillage on reducing direct runoff ranges from
slight to substantial. Procedures of the Soil Conservation Service used today
for estimating runoff do not consider the effects of conservation tillage and
no-till practices on runoff. To develop the SCS runoff curve numbers for these
practices, tillage and crop data were assembled from small watersheds and plots
under natural and simulated rainfall from many locations across the country.

The residue left on the ground was chosen as the independent variable to
represent the effects of conservation tillage practices. These data were stud-
ied to determine runoff curve numbers for single- and double-cropping systems
under various conservation tillage practices. These runoff curve numbers can
be used with the SCS procedure to evaluate the effect of tillage practices on
runoff.

INTRODUCTION

In 1976, about 17 million ha (10% of the farmed cropland in the United
States) were farmed with some form of conservation tillage (3J • SCS estimates
the use of conservation tillage, including no-till, has increased at an average
annual rate of 1.2 million ha over the past 10 yr and should continue to in-
crease because of its environmental benefits ( 3J • Conservation tillage is de-
fined as a form of noninversion tillage that retains protective amounts of re-
sidue mulch on the surface throughout the year (20). It includes such practic-
es as till plant, chisel plant, no-till, strip tillage, sweep tillage, stubble
mulching, chop plant, and other types of noninversion tillage. The relative
effectiveness of these practices in controlling runoff (_18) can be judged by
how much they:

(1) Reduce runoff velocity (JJ3, JJJ . The velocity of surface runoff wa-
ter is reduced by decreasing land slope or by increasing surface
roughness. Slope usually is decreased by lengthening the flow path
of the water. Surface roughness is increased by reducing the number
of tillage practices or by increasing vegetative or residue cover.
Decreasing runoff velocity usually increases infiltration.

(2) Increase surface storage. Practices that increase surface storage
generally reduce the total volume of runoff and increase infiltra-
tion.

\l Hydrologist, USDA-SEA-AR, Beltsville, Md.; agricultural engineer, USDA
-SEA-AR, Morris, Minn.; and h

Internet Archive Audio

Featured

Top

Images

Featured

Top

Software

Featured

Top

Books

Featured

Top

Video

Featured

Top

Mobile Apps

Browser Extensions

Archive-It Subscription

Save Page Now

Full text of "CREAMS : a field scale model for Chemicals, Runoff, and Erosion from Agricultural Management Systems"

See other formats